PRE2019 3 Group6

Group members

David van Son	1005864
Susanne Louvenberg	1238843
Jur Janssen	1247069
Bas Ohlen	0963529
Jeroen Meijs	1008703

Problem statement

In our current society the sitting position is the most frequent body posture, especially in office jobs. Many other professions (sometimes) require working behind a desk. Students also experience those working conditions. Jans, Proper, and Hildebrandt (2007) found that working adults in the Netherlands can spend up to 12 hours sitting down on a workday. Because people are sitting more hours a day, much research is done to determine the consequences of sitting for longer periods of time.

Research has been done about long-term health risk of long occupational sitting. Health risks, such as obesity, cancer, type 2 diabetes, cardio-vascular disease, and [mortality] are examined in their connection with occupational sitting. The authors of the paper concluded however that there is insufficient evidence of a causal relationship between those conditions.

Other research does show that occupational sitting increases pain. Medical and ergonomic field studies indicate that sitting posture can be the cause of [muscle, connective tissues of tendons, ligaments, and join capsules pain]. Chronic pain and other problems may be the result of static load for longer periods of time. The degree of pain increased as the time of occupational sitting increased. A study by Womersley, L and May, S (2006) showed that people with backache sat uninterrupted for longer periods of time compared to the people without backache . The sitting posture also determines the effects of occupational sitting. In their same study the group with postural backache also had a more flexed relaxed sitting posture. Other research confirms this result because slumped sitting position and poor shoulder posture (e.g. rounded shoulders, and head forward) causes pain due to mechanical changes that affect the function of the median nerve. Shoulder protraction reduces the nerve movement and other joints are moved. In response to moving other joints, the nerve dynamics is altered which changes the local blood supply. This is harmful for the nerve function and causes the risk of neck and shoulder pain.

Backache and neck pain are one of the most frequent cause of invalidity in industry in most Western countries. Kuoppala and colleagues (2008) showed in a systematic review that promoting ergonomics and a good sitting position reduces the absences from work. This stresses the importance of a good sitting position, because it reduces pain for individuals but also decreases work absences for the company.

Marshall, and Gyi (2010) mention: “Environmental influences such as a lack of support for the feet, low-friction seating material, or poor desk height can all create additional muscle work. Poor design forces the adoption of awkward and inefficient working postures that can ultimately lead to discomfort, pain, and chronic disability if adverse conditions persist.”. In addition to the environment influencing the sitting posture, another research paper states that individuals with neck pain have a different perception of a ‘good’ sitting position. Their sitting position is slightly different, and even a small change in head position can result in an increase of the lead on supporting structures and muscle activity. This indicates that it is important to impose a sitting position on people to accomplish a good sitting position that decreases the chances of pain.

To conclude, it is of importance to have a chair that provides a good sitting position to reduce the effects of occupational sitting. However, every person has a different physique, which means that one chair would not fulfil the needs of different users. Most chairs can be adjusted to some extent according to the users wishes. But as stated above, users who experience backache do not always have the correct idea of a ‘good’ sitting position. In the current working environment, employees do not have a fixed sitting position because of flex-work spaces. Therefore, the user needs to adjust the chair every day to have a good sitting position. To overcome all the problems stated above, this project envisions an automatic chair that helps the user with establishing a good sitting position.

Additional Papers

The following list consist of other papers that confirm the problem statement and are of relevance to this project.

• Posture plays an important role in performance. Poor posture can lead to worse task performance while also adding stress to the spine and balance muscles .

• Posture is also a tell-tale sign of engagement, it is even possible to estimate engagement purely on posture .

• This paper studied two groups, symptomatic and asymptomatic office workers. All subjects demonstrated an 10% increase in forward head posture from their relaxed sitting postures with the computer display. No substantial evidence for posture changing over a working day was found. .

• The high complain of musculoskeletal disorders is due to awkward postures, unsuitable workstation and lack of knowledge related to the areas to apply in everyday routine and it shows that working postures have a direct contribution on musculoskeletal disorders complained by the office workers in Putrajaya. .

• Given the association between RULA (Rapid Upper Limb Assessment) score and the prevalence of the problems, reducing RULA score by designing ergonomic workstation may reduce the prevalence of WMSDs (work-related musculoskeletal disorders) among the workers. .

• Computer usage increases risk of developing musculoskeletal disorders. Such an increase is mediated by ergonomic factors such as mouse use, remaining seated for prolonged periods, adoption of inadequate or uncomfortable postures, performing certain PC tasks, and psychosocial factors. .

Our Solution

To overcome all the problems stated above, this project envisions an automatic chair that helps the user with establishing a good sitting position. This chair has the possibility to automatically adjust the sitting position of the user. When the user wants to use the automatic chair, he or she needs to login. This is to know which user uses the chair and therefore which unique position chair needs to take. This can be done by scanning the user’s student or company card. Besides some other personal information, this card will have some details about your body part lengths. With this information, the automatic chair can adjust the sitting position for a particular user in the best sitting position to overcome backache.

Scenarios

The following two scenarios describe the importance of this project and the end user that is envisioned.

Fleur studies the bachelor Applied Mathematics at the TU/e. She needs to attend lectures and study for many hours a week. This means she spends about 5 hours a day on occupational sitting. Her days consist of meetings, lectures, and individual studying, which means she switches from different chairs very often. However, she does not take the time to adjust the chair to her optimal sitting position. Most of the time, Fleur only changes the height of the chair. But she started to notice that she is experiencing backache. She realizes this pain is coming from a bad sitting position. Therefore, she is enthusiastic about the new automatic chairs on the University. Since the new chairs arrived Fleur has been using the automatic chair every time, which is easy to use because of the login system. She is experiencing way less backache compared to before. The chair made it easier to adjust the seating position which she didn't completely do before. Moreover, the automatic chair made her more aware of her sitting position.

Thomas is a 56-year-old who has already been working for Rabobank for 30 years. His job requires him to work behind a computer every day. He experiences occupational sitting for around 8 hours a day. In the past, Thomas experienced shoulder and neck pain. However, he searched for help and understood it was because of the many hours sitting in a bad position. From then on, he started to adjust his chair as much as possible to have a better sitting position. He has been doing this for almost 10 years already. As a result, he experiences far less backache than before. But a few years ago, the Rabobank started to use flexible working spaces. Which means Thomas needs to switch places every day. This is very inconvenient for him, because he needs to adjust the chair again each day. Because he does this in the most optimal way, it takes him 5 minutes each time. Thomas would really like to see the automatic chair in his office. This would mean not having to struggle each day to adjust his chair.

Objectives

There are many different kinds of users that could use our product. However, we consider only the primary user as our focus for this project. A primary user is someone that experiences occupational sitting in a flex environment for long periods of time. Examples of this are office workers or students.

Primary user objectives

• The user can ‘activate’ the chair to automatically go in the good sitting position.
• The user can manually change the sitting position of the chair.
• The user can use the chair like a regular chair, and thus without the automatic option.
• The user understands what the sitting cues indicate.
• The user knows how to sit in the correct good sitting position.
• The user can adjust their incorrect sitting position to the defined good sitting position with the given cues.

Survey

To find out more about what the primary user would want from our product, a survey was done. In this survey we examined the primary user's perspective on our product. To get accurate results we choose to do the survey among students because they are part of the primary user group.

The main goal of this survey was to find out whether the user would like to use such an automatic chair to begin with, and what requirements the user would have for our project. It is important to get an idea of the current situation and the preferences of the user, because this helps us to steer the project in the right direction.

This survey consisted of questions regarding the users sitting behavior and about our solution to a bad sitting position. It can be divided into three parts based on the questions that were asked. The first part are questions about the current situation. How people adjust their seats and sit on them. The second part asks what the participant would want in a chair. The third part asks what our product could do to help them sit in a good position, and what the user would want from our product specifically.

The survey is linked here.[1]

Results

The results are linked here.[2]

Questions about the current situation

In the survey, it was asked whether users change workplace often or not. The participants gave mixed answers. Some switched chairs often, but most people do not seem to switch chairs that much. It was also asked whether the participant is aware of their sitting position. Only half of the people indicated to be aware of this. But only a third actively changes their bad sitting position. 40% of the users said they adjusted their seat often, another 40% said they did not adjust their seat at all, and the remaining 20% said they adjust it sometimes. The height of the chair was mostly adjusted. The armrest was the second most important and the back of the chair coming in third. When people adjust their seat, only a few of them spend a minute or more on the adjustment. The overwhelming majority spends much less than that.

Questions about what they would want

The survey showed that people would prefer to have their own personalized chair. Logically, people preferred an adjustable chair over a nonadjustable chair. Participants wanted to be able to adjust their chair. Height was most important followed by the arm rests and the back rest.

Questions specifically on the automatic chair

It was asked if people would be fine with a server that keeps data on users’ preferences and their location. Almost all participants replied that they would be fine with that. A few of them has some concerns about privacy and would only allow it if the data was anonymous. Half of the participants found an automatic system convenient, and the other half also saw the benefits of a good sitting position. After the automatic seat adjustment, almost all people want to have the option to adjust their seat manually. Most people would want to make use of the automatic chair.

Conclusions

When looking at how people sit, it can be concluded that users are not very concerned about their sitting position. Most people do not seem to actively do something about their position. A reason can be that users are most concerned with comfort rather than the effects of occupational sitting. The answer on which sitting position they have, as well as the fact that more users adjust their seat than care for it, support this.

When adjusting, the most important thing is the height of the seat, followed by the armrests and the backrest. When designing our product, these will be the most important parts of the chair that need to be changeable.

It can be concluded that convenience is an important factor that should be taken into account in designing the automatic chair. This is supported by the fact that all of the participants liked the convenience of an automatic system.

Our Goal

The survey provides some important remarks about the automatic chair. It stands out that users want to manually adjust the chair after it has been automatically changed into the good sitting position. Users mentioned this is important because: ‘the position of the chair can be experienced as not comfortable’. Another questioned showed that users have many different sitting positions. Of which not many people follow the backrest.

This started a thought process. Engineers can design the best automatic chair which will change to the perfect sitting position, but can it be assumed that the user will sit on this chair with a good sitting position? Probably not. The results of the survey show that users will adjust the chair and/ or will sit in a relaxed bended position. The goal of the automatic chair is to reduce the pain caused by occupational sitting. The automatic chair is designed to provide the user with a chair that helps to sit in a good position. However, the next step is to make sure the user uses the chair as intended and stays in this position.

As can be read in the problem statement, pain caused by a bad sitting position is common and can be reduced by accomplishing a good sitting position. Reducing occupational sitting pain is our main objective. There is already done some research about the systems that make the automatic chair. This is described in the section State-of-the-Art. Because of that, the focus of this project shifts into the second design step of making sure the user stays in this good sitting position.This includes registering the users sitting position on the chair, knowing whether this is a good sitting position, and finally giving a cue to the user if this is not a good sitting position.

Product Requirements

The following list of requirements help to structure the project process and ensures traceability. These are the requirements for the prototype. This includes a pressure sensor mat that measures your sitting position. But also, some type of cue that will let the user know when they have a bad sitting position. Other requirements would need to be added, if the list was about the whole automatic chair.

In order to work for most people, the chair and sensors need to be able to hold the weight of most people.

• The sensors need to hold up to a weight of at least 95 percent of the population (97.98 Kg).

To make sure users will actually use the automated chair, it should be comfortable. Otherwise users will simply choose another chair to sit on.

• The sensors and cues should not interfere with the comfort of the user.

The cues provided by the automated chair should be clear, this enables the user to better understand what they are supposed to do. This increases comfort and ease of use, which would increase the use of the chair.

• The cues need to be clear so that the user knows what they mean after explaining it once (9 out of 10 users).

If users do not find the chair useful or comfortable, they will not use the automated chair, so the chair should be more desirable to use than a regular chair.

• The prototype should be desirable to use, so that the user would rather sit on the automatic chair than on a normal chair (9 out of 10 users).

The chair is only useful if the user is willing to change their position, so the chair has to make the user want to change position after giving the cue that they should change their position.

• The prototype should make the people change their sitting positions after the cue is given (9 out of 10 users).

We choose to require 9 out of 10 users for most requirements, because this technology is designed to become very commonplace, so most, if not all, users should be accommodated in the requirements, and thus the design.

State-of-the-Art

The automatic chair can be build based on three systems. Firstly, a log in system in the chair which will provide the information about the user that is using the chair. Secondly, the system needs to know what a good sitting position is for this particular user. Thirdly, this information will be used by the system to automatically change the position of the chair. There already has been done research about all systems, separately or combined. Here an overview is provided about current research relevant to the project.

Memory device for a user profile

For the automatic chair a login system is envisioned to make it easy to adjust the chair for a particular user. Such system can already be found in many technologies, as for example: TU/e printers, AH bonus card, and Android Multiple User. Another example is the memory device for a user profile of devices in a motor vehicle . This innovation: “is used for providing data corresponding to the user profile in the vehicle without a user having to make corresponding settings.” This memory device is useful because it consists of personal data but also activation data for the vehicle. In the case of this project, the vehicle could be the chair. This memory device may be used independently of a vehicle and is therefore very flexible to use.

Ergonomic guidelines for a chair

Much research has been done about what the good sitting position is. A paper by Zheng, Dorsey and Miltra (2014) describe the ergonomic guidelines for an ergonomic chair. . Those guidelines are listed below.

• The seat of the chair should have the correct height. Both feet should be supported. When a chair is too high, it creates undue pressure at the knee and thigh. While, if it is too short the knee will be higher than the hip sockets.
• Width and depth of chair seats should conform to the user’s dimensions.
• Flat un-contoured seats are preferred to discourage a slouched or C-shaped posture.
• Lumbar support by providing low- or mid-back support can help hold good posture and prevent pain to the spine and neck.
• Head support, if provided, can help ease stress for the neck muscles and provide support for seating over extended periods.
• Arm rests provide support for reading, typing, painting, and similar activities.

Research on ergonomic design and evaluation of office backrest curve

This other paper forms a good basis to establish a chair with a good sitting position . This paper conducted a survey which gave interesting insights. Results showed that the most used sitting posture is the ‘relax’ posture seen in Figure ??. The survey shows that 50.3% of the subjects considers the backrest as very important. As part of the backrest, the waist support causes pain in the back when sitting in an office chair for a longer period for 58.09% of the participants. Followed by the neck support part of the backrest which causes for 57.23% of the participants pain in the neck. The backrest inclination angle (36.01%) and the hardness/ softness (31.83%) of the backrest are also causing discomfort. Thus, when the back of the user cannot fit well in the backrest due to shape and material, it eventually will cause neck and shoulder pain. This paper concludes that the backrest is the most important part of an office chair.

Sitting posture most commonly used by employees

A test was done to see whether the shape of an office chair corresponds to the shape of a spine. The results of the chair and spine measurements can be seen below. The shape of the spine of an average person can be seen, together with the shape of the four tested chairs. This shape is divided into three parts: head and neck, back and thirdly the waist. None of the four chairs is similar to the human spine when sitting upright. All the chairs do have a waist support, but not fully consistent. The most serious differences are at the head and neck area, but also the upper back is not well supported. This shows us that most of the existing office chairs do not follow the shape of the human spine. This research also showed that the chair backrest is mainly used for relaxation. It plays a small role in supporting the user while working. It is suggested to design the back of the chair according to the shape of the human spine to support the human body while working. It is of importance to match the curve of the back of the chair with the shape of the spine in the sitting position.

Construct of spine and chair backrest

This paper concluded with the following ergonomic requirements for an office chair.

• Headrest height: 628.3 – 675.1 mm. (ranging from the normal height for females to that of males with high cervical spine point in a sitting posture).
• Waist support height: ≥ 210 mm.
• Waist support depth: 20 – 40 mm.
• Effective back width: ≥ 360 mm.
• Seat back height: ≥ 460 mm.

Active approach to improve ergonomics

In this paper an active approach is made to improve ergonomics by combining sensing and self-actuating workspace furniture . Posture sensing, ergonomics reminders, and active furniture were combined to improve ergonomics. Possible options of posture sensing are:

• Accelerometers in wearable devices that can track partial body postures.
• Flex sensors that can detect head tilt and arm angles.
• Capacity sensors and piezoelectric sensor used in chairs that can detect bad postures on pressure distribution.
• Vision-based monitoring systems that can detect sitting postures.
• Geometric features can determine incline angle of user’s head.
• Face detection that can calculate the distance between face and screen.
• Microsoft Kinect sensors that can provide skeletal tracking, measuring the user’s body dimensions.

This study made a prototype of an automatic chair including ergonomic reminders. There is a real-time feedback displayed on the screen. This guides users on how to adjust the chair position and height for a good sitting position. This product is an active furniture that uses a motorized desk for automated height adjustments. In addition, dual robotic arms provide automated adjustment based on sensor data on height and distance. The ergonomic guidelines that were used are:

• Maximum forward head tilt of 15°. (1)
• Upper arms are vertical, and forearms are horizontal. (3, 4)
• Thighs are horizontal, and knees are at 90°. (5, 6)
• Vertical viewing angle of 15-20° below the horizontal, with the first line on the screen below eye level. (2)

The 6 posture angles that require personalization in a computer workspace

The paper mentions that only a prototype was made and it needed improvements to become a real product. One suggestion as future work was to conduct an extended field study. This would be needed to observe the deviation from the initial postures. We also envision an active approach supporting continuous posture and activity monitoring for helping users maintain ergonomic postures throughout the day.

Automatically adjustable office and task chairs

Already back in 1996, Google placed a patent on their designed automatically adjustable chair . Their innovation is capable of five electrically powered position adjustments. But in all cases, the user is able to adjust the chair himself. This is: “to reduce the strain of sitting in exactly the same position of extended periods of time and reduce repetitive motion injuries”. This chair also has a memory device. This way multiple users can quickly adjust the chair to a preselected position. The idea behind is to make minor adjustments to the users position over periods of time to again reduce the strain.

A system for posture monitoring and guidance

This article has as goal to improve the sitting behaviour of office workers . People are usually not aware of their (statically) sitting behaviour and posture while working concentrated on a task. An intelligent office chair was designed that measures the users sitting posture and whether this person sits statically. With effective feedback, the chair will guide the user with sitting in a more dynamic and healthy way. The chair is equipped with four force transducers, that detect the sitting posture of the office worker. If an ‘unhealthy’ posture is detected, an alert directs the user to sit in another way.

This research focuses to detect a certain posture but mainly to estimate the biomechanical effect a certain sitting posture. It turns out that there are two important parameters: “the time a person sits statically, and another important parameter is the force acting on the spine which is closely related to the flexion and extension angle of the lumbar spine”.

An office chair was equipped with four force transducers, located at each corner under the seating. These four sensors make it possible to compute the center of Pressure (COP). “Only four force transducers are required to estimate characteristic parameters for quantifying the biomechanical effect of a certain sitting posture.” This allows to monitor the sitting position of an office worker.

The 4 sensors used positioned on the chair to determine the Center of Pressure

Where We Continue

As can be read above, there is already quite some knowledge that is needed to make the automatic chair. Because of that, this project assumes an automatic chair that is envisioned can be designed and produced. To continue the research, our focus will be the second design step of making sure the user stays in a good sitting position in this automatic chair.

Our research question is: How can the user be stimulated to stay in a good sitting position indicated by the automatic chair.

Possible Solutions

There are many different ways the user could be stimulated to keep a good sitting position. The possible solutions mainly differ in the amount of autonomy of the user. One solution could be to launch an information campaign that raises awareness of the problem. This option leaves the user with the most autonomy. In this case, the user would be able to decide for himself whether he actively adjust his behavior because of the information. Another solution which involves a seating police limits the autonomous decisions of the user way more. Imagine a scenario where citizens monitor each other. In our case, this seating police would consist of many normal users, who could watch others whether they are seating like they are supposed to. A warning or punishment could be given to force people to sit in a healthy way. While these solutions are on the ends of the spectrum of user autonomy, there are also more balanced options. These solutions came down to warning the user of their bad seating position, either actively or passively. This way the user remains their autonomy for the most part, while being nudged in the direction of a healthy seating position. The following list consists of possible solutions.

Informing

• Raise awareness and informing people about a healthy seating position.
• Measure the current way the user sits and give information on how to improve.

Nudging

• Built a display in the chair which shows if you have the correct sitting position.
• Built a light in the chair which shows if you have a correct sitting position.
• Notification on your phone which reminds you of your sitting position.
• Let the chair vibrate if the user does not have a good sitting position.

Paternalism

• An auditory stimulus to let the user know it should keep the good sitting position (similar to seatbelts in a car).
• A blocking system on your computer that only allows the user to use the computer when it has a good sitting behavior.
• A seating police.

Conclusion

Based on the survey held in week 2, it can be seen that users are already aware of their bad sitting position. This indicates that a lack of awareness is not the problem. Therefore, an information campaign would have little to no effect to solve the bad sitting position of users. It was also already mentioned that users would like to always have the possibility to adjust the chair, which indicates that they value their autonomy. Those users would probably not like to sit on an actively warning chair. If the encouragement for keeping a good sitting position is to annoying and/ or frustrating for the user, the user would probably sit somewhere else. This will most likely result in them sitting unhealthily, which is opposite to the goal. These findings point out that a chair which encourages the user to keep a good sitting position would be best. A solution which involves passively warning the user would be most suitable.

Measuring The Sitting Position

To passively warn the user of their bad sitting position, it is required to measure the sitting position of the user. This can be done by using a pressure sensor map. This mat can be put on the office chair. This way it will not be needed to make a whole new chair. Inside the pressure mat sensors will be placed equally divided in a 3 times 3 structure. Andreas Schrempf et all. (2011) have shown that the sitting position can be determined by using 4 pressure sensors. However, it is also that the more sensors the more precise the sitting position can be determined. Because of that, this prototype consist of 9 sensors.

Pressure sensors

Load cells

Load cells have different kind of ranges, there are loads cells that have a range of 5-10 kg (55mm x 12.7mm x 12.7mm), but also up to 200 kg (150mm x 38mm x 24mm). The working of a load cell is not ideal for our situation. On both sided (up and front) a piece should be mounted. Then a force will be applied on one plate than the straight bar will deform and based on the deformation the pressure can be translated into an electrical signal. The price of load cells ranges from €10 - €15. [[3]]

Force transducers

Force transducers are used for dynamic, short-duration static and impact force measurements. It can measure tensile and compressive forces, this can be option for our product. The maximum compression is about 80kg, this enough because the transducers will be divided over the whole seat. The dimensions are 19.05mm x 15.93mm. [[4]]

Force sensing resistor

A force sensing resistor (or force sensitive resistor, FSR) is a material whose resistance changes when a force, pressure or stress is applied. These FSR’s have a maximum range of 10kg, it is not sure if this is enough. The weight of a person will be divided over the whole seat, so if enough resistors are used than 10kg can be enough. There are different kinds of resistors. A square FSR (44x38mm) of €9.95 or a circular (12.5mm) of €6.95. [[5]] There is also another one, this one is much more expensive, €21.95. But this one has a much bigger range because the resistance can be adjusted, the maximum can be set up to 300 kg. [[6]]

Conclusion

The choice will be between the circular resistor with a diameter of 12.5 mm or the square resistor with the dimensions 44x38mm. For this project it is better to pick the bigger sensing resistor, because the image of the pressure distribution will be the clearest.

Pressure mat material

The material for the pressure mat is also important. This material must be comfortable because a person will sit on the mat. Also, the FSR's will be places in this mat. It is a requirement that the FSR’s will not be felt by the user, and thus must the material be able to fulfill this requirement.

Polyether SG35 or SG40

This material is often used for seat cushions and the hardness is medium. The material is very cheap for a 300x400x30mm piece the price is €1.80.

Koudschuim HR40

This material is often used for chairs and matrasses. The material is very cheap for a 300x400x40mm piece the price is €3.00. Minimum height is of this material is 40mm.

Conclusion

It does not really matter which of these two products is chosen for the prototype. Both materials are sufficient. Polyether SG40 is chosen for this project. Two thin mats can be bought to place the FSR's between those mats.

List of Materials

Product	Quantity	Price
Force Sensing Resistors	9	€114.35
Foam, SG40		€18.43
Luidsprekerkabel	25m	€15.85
Tape
Arduino Mega 2560	1
USB/USB-A cable	1
HEF4051BP Analog Multiplexer	3
Vibrating engine	1	€0,99
Resistor 10 kOhm	6

Prototype

Building the Prototype

The materials listed above were ordered. It was important to first have the mats to place the pressure sensors in between. The polyether SG40 was cut into the mat of 50x45 cm. This was around the average size of a seat. Then, the positions of the resistors where marked. The pressure sensors were placed in a 3x3 matrix with the same length between each sensor. To connect the pressure sensors with the Arduino, ‘Luidsprekerkabel’ is used. This cable is cut into the right length. At this moment, the wires are soldered to the pressure sensor resistors. Finally, the sensors are placed on the mat with double-sided tape. However, this tape does not stick very well on the material of the mat. It was found that regular duct tape works better, but not very well. To make sure the wires stay at the same place, cuts were made in the foam mat. This way, the wires are placed into the mat which was very stable.

Prototype

Position detection

The automated chair needs to be able to detect whether the user is sitting in the corrent position. A good estimation of the position can be extracted from the center of mass. This is also described in the state of the art . So, we choose to use the center of mass as our indicator that the user is sitting correctly. If the center of mass is in the center of the seat, the user is sitting correctly. If the center of mass is towards the edges of the seat, the user is sitting incorrectly.

This approach simplifies the position detection a lot, but also neglects a couple of things. For example, the position of the legs is not measured, also, the position of the head and arms cannot be measured in this way. However, this method does provide a lot of information about the position of the back, which is central to a proper posture.

How does it work

The prototype consists of the the circuitry, the Arduino, the actuators and a PC.

Circuit

The circuit diagram is shown in [7]. The circuit makes use of a voltage divider structure. This divides the voltage between the pressure sensor and a (static) resistor. Since the pressure sensor's resistance decreases when the force applied to it increases, the voltage across it also decreases. Since the total voltage across the pressure sensor and the resistor combined remains the same, the voltage across the static resistor increases.

To be able to use a large amount of sensors, analog multiplexers are used. These components can connect multiple inputs to one output, by selecting only one input at a time to be connected to the output. The multiplexers are controlled by the Arduino.

Arduino

The Arduino is responsible for three things, controlling the multiplexers, driving the actuators, and converting the voltages across the resistors to bytes, which it sends to the PC. The voltages are converted by the Arduinos internal analog-to-digital converters (ADC). The ADCs have a resolution of 10 bits, but we only use 8, since that is faster and more precision is not necessary.

The multiplexers are controller via three pins, these pins represent a binary number, which tells the multiplexer which input should be connected to the output. The pins are set using a high (5V) voltage for a '1', and a low (0V) voltage for a '0'. The bytes are sent to the PC via serial communication (baud rate = 115200).

The actuators are driven by setting the output pin of the Arduino to high (5V), this then powers the actuators.

PC

The PC gets as its input the bytes sent by the Arduino over serial communication. The PC then visualises these bytes as a shade between white and black, representing pressure points on a pressure map. It also calculates the center of mass as a weighted average of the pressure points. This center of mass is also displayed by the PC in the pressure map.

When the PC detects that the center of mass is outside the boundaries, it tells the Arduino to drive the actuators.

Software

General overview

2 separate pieces of software were needed for our project. The first piece of software controls the Arduino and sensor mat itself. The second piece of software runs on the laptop and is in charge of controlling the Arduino and visualizing the read sensor values. The Arduino software is written in C. As that is the default language for programming Arduino. The Arduino IDE was used to write the code. The laptop software was written in C++ with Visual Studio as editor.

Arduino

The software on the Arduino is needed for reading of the sensors and sending it to the laptop. The sensors are connected directly to the inputs of the Arduino. The multiplexors are connected to the output ports of the Arduino. The software reads the data of all the sensors in multiple steps. First it selects the sensor row by setting the correct value for the multiplexors by writing to the output pins. Then it will read the sensor data. It will repeat this process until it has read all rows. The raw sensor data can be a bit sensitive to small fluctuations. That is why an average is taken of the current measurement and previous ones.

The values of the filtered data is collected as an array which is then send over the serial bus to the connected laptop. The sampling rate of the Arduino is a few kilohertz. That means that the Arduino can deliver measurements faster than we need. For displaying we only need about 60 samples per second. Any more cannot be visualized.

We have set up the communication between the laptop and the Arduino such that the laptop can poll the Arduino for a current reading of the preassure mat. The default state of the Arduino is there for listening to the serial bus for a read command. Only when it receives such command it wil actually read and send the sensor data.

Laptop

The laptop does the most important work. Besides visualizing the data it is also responsible for all processing of it. The laptop software has an internal clock. Each step it sends a requenst to the Arduino for a current measurement. It then listens to the serial bus for the received data. The laptop will the process the data. It will interpret the array as a 2D grid. It will calculate the center off mass.

There are many ways to calculate a weighted average or center of gravity. We choose to calculate it by treating each sensor as a particle with a position and a certain weight. The heigher the measured pressure the more the weight of the particle is. The particles are arranged in a grid. The center of mass is then calculated for this particle system <insert formula>. This method works quite well in the test we have done. A downside is that the sensors should be spaced equally. This method does not compensate for sensor density.

The sensor map is also visualized as a height map. Bilinear interpolation is used to smoothen out the image. Using OpenGL this is then displayed as a height map with colors based on the height values. The center of mass is also overlayed on top.

Another important task of the software is detecting if the user sits in an unhealthy position. For our approach we have defined boundaries for our center of mass. If it is outside those boundaries for too long it indicates that the center of mass of the sitting user has shifted and he/she is likely sitting in a ‘wrong’ position.

As soon as this is detected a signal can be sent to the Arduino the turn on the vibrating motor in the chair or to turn on a light. When the center of mass return to within the boundaries the user is sitting correctly again and the signal is reset. For the motors the laptop will send a signal to turn off the motor or light.

Experimental plan

Introduction:

To keep the user in the design process of the smart chair system, different experiments are done to determine the most optimal settings. The experiments are based on a study done for the National Highway Traffic Safety Administration of the US. In this article the effectiveness of seatbelt reminders is tested. There are 5 different settings that are tested (basic reminder, continuous flashing, periodic reminder, aggressive reminder, one long reminder) each of these approaches is tested on three different parts. During the experiment the effectiveness, annoyance and attention getting of the signal setting is rated by the test subject and after the experiment a questionnaire is filled in for effectiveness desirability and preference.

At the end of the experiment the participants commented on different signals. They found that the best way to give a visual reminder is a system that gets progressively brighter or flashes increasingly over time. The best acoustic signal is a voice message that comes on periodically and a close second is a non-voice noise that does the same thing. The participants also desire a way to customize the signal to their own preference.

The conclusion of the experiment was that the more annoying the signal was the more effective the response of the participant. Also the desirability of each system in relation to annoyance is different for each participant, some favour the more annoying systems while others desire the more nuanced system. The use of only a visual signal is not effective and should always be supported by an auditory signal. A visual signal should always be flashing because a static visual signal will not attract attention. The main difference in this study in comparison with our system is that the signal should not be so annoying that the user is not willing to use the chair. One of the requirements states that 9 out of the 10 users find the cues not to annoying and uses the chair voluntarily. This is something is experimented in this user test.

The first experiment is to find what kind of signal will give the best results. The second is a qualitative study of the system, where we ask for the opinion and remarks of the user to further improve the settings.

Experiment 1

Goal of the experiment

The goal of this experiment is to find out what is the best way to let the user (the person sitting on the chair) know that they need to change their sitting position.

Description of experiment

The participants take a seat on a chair with the prototype placed on it in front of a laptop, they are informed on what the prototype does and that a cue would inform the user that they should reposition. To simulate a work environment they are asked to fill in a simple sudoku during the experiment, this is so they are focused on the computer as they would be in normal conditions. Then different methods of signalling are tested and the participants are asked to rate each method.

1. Vibrating of the chair in regular intervals of 20 seconds

2. A constant vibrating of the chair

3. A blinking light

4. A constant light

5. An acoustic signal in regular intervals of 20 seconds

6. With a pressure map of the chair

All four of the signals will be tested and the user will grade the signal on a scale of 1 to 5 for each statement, as can be seen below. The experiment will be repeated to get a confident result.

At the end of the experiment the participants are asked if they have any remarks or idea’s on the signal procedure.

Results Experiment 1

The results will be collected and the best way will be chosen. The signal has to be both effective and desirable, because it is wanted that the users will sit in the correct way but not that it is too annoying that the users will not use the chair. The best result will be used in the next experiment. The experiment is done by 5 participants. The average scores are given in the tables below for each signal. Sometimes there are two crosses for one question, this is because the average is in between two answers (e.g. between Agree Slightly and Agree). The boxes are filled with colors to get a clear overview of what is the best signal to use.

Signal 1: Vibrating of the chair in regular intervals of 20 seconds

Signal 2: A constant vibrating of the chair

Signal 3: A blinking light

Signal 4: A constant light

Signal 5: An acoustic signal in regular intervals of 20 seconds

Signal 6: A pressure map of the chair

Conclusion

Looking at the tables, 2 two signals will not be used because these are: a constant vibrating of the chair, and an acoustic signal in regular intervals of 20 seconds. These two signal are too annoying and people want to turn this signal off if they can, so than the use of the signal is gone. The signal of a constant light will also not be used, it was not really clear to the participants that a signal was given. Then there are three signals left. The two best are a vibration with a regular interval and a pressure map of the chair. So these two signals will be used for our prototype and for the second experiment.

Experiment 2

Goal of the experiment

With a better idea of what kind of signal to use to get the best results, the experiment continues with a qualitative study. This means the users use the system and ask for their opinion and remarks for further improvements. This way we can optimize the timing, duration and design of the system.

Description of experiment

The prototype is placed in normal working environment (flex workplace, library, etc.), and people are asked to sit on the prototype while working or studying. The signal will go off when they need to readjust their posture. After 15 minutes, the participants are asked to answer the questions that are stated below.

1. What did you think of the sensitivity of the system (the time between when the user sits in the wrong position till the signal is given)?

2. What did you think of the duration of the signal?

3. What did you think of the intensity of the signal?

4. What did you think of the comfort of the prototype?

5. Are you willing to use the system if it is properly introduced?

6. Do you want to be able to turn the signal on and off?

7. What did you think of the visualisation feedback?

8. What do you want to change in the system?

9. What do you like in the system?

10. Do you have any further ideas or improvements to the system?

Results

The results will be collected and the system settings will be tweaked to the needs of the users.

Planning

Below a small week to week planning for the project:

• Week 4: Finish the programming and electrical scheme, order all the parts for the prototype.
• week 5: Build and test the prototype.
• week 6: Do the experiments and collect results.
• week 7: Process the results and begin on the presentation.
• week 8: Finish and prepare thepresentation

Approach

Our approach is that we start by gathering information regarding our topic, the state of the art and the relevance of our research. We will then hold a survey among people who use adjustable chairs often, in which we want to find out which part(s) of the chair they most often adjust. Using this data, we will research which parts are in most need of being monitored. Then we will determine possible ways of warning the user, and make prototype(s) of these systems. We will then test which way is preferred by the user, and which way gives the best results. Combining these results, we will conclude which way would be best for a user warning system.

Milestones

• Evaluation of the best working posture.
• Made and held the survey
• Determined the most relevant adjustable parts of a chair
• Determined the sensors that are needed to detect a person’s working posture.
• Made a prototype of the user warning system
• Full test evaluation of the user warning system
• User evaluation of the user warning system

Deliverables

• This Wiki page containing all our research and findings.
• Survey results about the adjustable chair.
• A prototype of the user warning system.
• Test and user evaluation of the user warning system.
• A presentation at the end of the project.

Who is doing what

Week 1

Name	Time spent	Break-down
David	11 h	Introductory lecture (2h), Brainstorm (1h), Studied papers (4h), Wrote summary (1h), Group meeting (2h), formatting wiki page (1h)
Jur	10 h	Introductory lecture (2h), Group meeting (2h), Studied papers [7-10] and made summary (4h), Brainstorm about possible topics (1h), Approach/Milestones/Deliverables (1h)
Jeroen	9 h	Introductory lecture (2h), Group meeting+brainstorm (2.5h), Studied papers(4h), Made user requirements (0.5h)
Bas	9 h	Introductory lecture (2h), Group meeting (2h), Brainstorm (1h), Studied papers, Update wiki(4h),
Susanne	10.5 h	Introductory lecture (2h), Brainstorm (0.5h), Group meeting (2h), Studied papers (2h), Wrote problem statement (4h)

Week 2

Name	Time spent	Break-down
David	8.5h	Tutor meeting (0.5h), Group meeting1 (1.5), rewrote approach, milestones and deliverables (2h), Group meeting2 (1.5h), Enquête (2h), data analysis (1h)
Jur	12h	Tutor meeting (0.5h), Group meeting (1.5h), [Search papers, summarize, make ready for Wiki, put on Wiki] (8h), enquête (2h)
Jeroen	12h	Tutor meeting (0.5h), Group meeting 2x (3h), research in ergonomics (4h), make a script for optimal position of the chair(4.5h),
Bas	7h	Tutor meeting (0.5h), Group meeting1 (1.5h), Group meeting2 (1.5h), enquête (2h), Made enquête (1.5h)
Susanne	10.5h	Tutor meeting (0.5h), Group meeting1 (1.5), Made enquête (1h), Group meeting2 (1.5h), Enquête (2h), Wrote objectives and requirements (3h), Wrote our solution (1h)

Week 3

Name	Time spent	Break-down
David	18h	Group meeting1 (1.5h), data analysis of survey (2h), add survey to wiki (0.5h), rewrite approach, milestones and delivarables (0.5h), Group meeting2 (1.5h), design electric circuit (2h), make prototype of circuit and program arduino (8h), write code documentation (2h)
Jur	11h	Group meetings 1 and 2 (3h), Tutor meeting (0.5h), Research sensors -> what is already used -> which is the best for us (5h), Research material mat (1h), Write parts for sensor and mat on the Wiki (1.5h)
Jeroen	10h	Group meetings 1 and 2 (3h), Tutor meeting (0.5h), Research into different sensors to use (6h)
Bas	11h	Group meetings 1 and 2 (3h), Tutor meeting (0.5h), Brainstorm (2.5h), Survey results (5h)
Susanne	13h	Tutor meeting (0.5h), Group meeting1 (1.5h), Rewrite our solution, Rewrite requirements (1h), Write two scenarios (1h), Group meeting2 (1.5h), Rewrite state-of-the-art and add papers (4h), Write our goal (1h), Write where we continue (2h), Upload wiki and change reading order (0.5h)

Week 4

Name	Time spent	Break-down
David	10.5h	Tutor meeting (0.5h), Group meeting1 (1.5h), Group meeting2 (1.5h), Update visualisation software (6h), write README document (1h)
Jur	7.5h	Tutor meeting (0.5h), 2x Group meeting (3h), Find and order materials (2h), Update wiki page (conclusions and material list) (2h)
Jeroen	7.5 h	Tutor meeting (0.5h), 2x Group meeting (3h), start on experimental plan (4h),
Bas	h
Susanne	8h	Tutor meeting (0.5h), Group meeting1 (1.5h), Group meeting2 (1.5h), Add research papers and finish state-of-the-art (4.5h)

Week 5

Name	Time spent	Break-down
David	10.5h	Tutor meeting (0.5h), Group meeting1 (1h), Working on prototype (5.5h), Group meeting2 (1.5h), work on prototype software (2h)
Jur	h	Tutor meeting (0.5h), Group meeting1 (1h), Working on prototype (3.5h), Group meeting2 (1.5h), Write prototype part (1h),
Jeroen	9h	Tutor meeting (0.5h), Group meeting1 (1h), do research in annoying signals (2.5h), Rewrite experimental plan(2.5h), preperation for experiment(2.5h),
Bas	h
Susanne	10h	Tutor meeting (0.5h), Group meeting1 (1h), Working on prototype (5.5h), Rewrite objectives (0.5h), Rewrite requirements (1h), Group meeting2 (1.5h), Add Tutor Meeting Questions

Week 6

Name	Time spent	Break-down
David	14.5h	Group meeting (1h), Finish prototype (10h), User test (1.5h), Test and debug latest software (2h)
Jur	h	Group meeting (1h), work on the Wiki (2h)
Jeroen	4h	Group meeting (1h), work on wiki(2h), prepare experiment(1h)
Bas	h
Susanne	7h	Group meeting (1h), Adjustment to requirements (0.5h), User test (1.5h), Read and make adjustments to the whole wiki page (2.5h), Add new article in state-of-the-art (1.5h), Add Tutor Meeting Questions

Week 7

Name	Time spent	Break-down
David	h	Group meeting, update material list, update circuit diagram
Jur	h	Group meeting (1h)
Jeroen	4.5h	Group meeting (1h), work on requirements (1h), user test (2.5h)
Bas	h
Susanne	6.5h	Group meeting (1h), Make adjustments of how results of first survey are shown on wiki page (0.5h), Make a setup / text of the presentation (5h), Add Tutor Meeting Questions

Week 8

Name	Time spent	Break-down
David	h	Group meeting, update requirements, update prototype, write discussion, write further research.
Jur	h
Jeroen	h
Bas	h
Susanne	h