PRE2017 4 Groep8

Group members

• Isabelle Cooijmans | 1014516 | i.h.m.cooijmans@student.tue.nl
• Ramon Hameleers | 0998964 | r.j.e.hameleers@student.tue.nl
• Angelique Husson | 0956648 | a.c.e.husson@student.tue.nl
• Marrit Jen Hong Li | 0963568 | m.i.jen.hong.li@student.tue.nl
• Dana de Vreede | 1020836 | d.d.vreede@student.tue.nl

Previous idea

Our Previous Idea can be found on this page.

Introduction

Problem statement

Elderly women with dementia often have problems managing their daily routine and thus independent living. By creating a system which helps with creating a daily routine the general health and independence can be improved.

Objective

Our goal is to help elderly women in the early stages of dementia. They will be helped in the form of an interactive system primarily focussed on letting the patient water its plants at specific times and let the plant be a reminder for certain periodical things, like eating and medicine. Firstly, by creating this repeated event, elderly will create a stronger daily routine. Such a routine provides consistency and a predictable time slot, in which to return to valued occupations . Secondly, our objective is to research whether the shape of such a system contributes to the functionality of it.

Thus the main objective can be split into two sub-objectives:

• A system that creates a routine for the patient.
• Investigating the impact of shape on functionality.

These sub-objectives will be explained in more detail further in this wiki page.

Hypothesis

The shape of a plant helps the user better than the shape of a cube.

Users

The users of the technology formulate certain requirements or wishes for the functioning of the technology. The users are a diverse group and therefore they will be split up into three kinds of users: primary, secondary and tertiary users. The primary users are the users that come directly into contact with the technology and directly benefit from it. Secondary users will use the technology infrequently or not directly. Tertiary users are the users who are affected by the technology or make decisions about its purchase.

The primary users are:

• Elder women in the beginning stages of dementia

The secondary users are:

• The care takers
• The family, friends and loved ones
• The government

The tertiary users are:

• The technicians

Requirements

The three different types of users described in the previous section all have requirements. These requirements are listed below. Requirements of primary users:

• Elder women with a decline in memory need a system that will help them remember specific tasks
• Elder women need a daily routine to improve their well-being

The specific tasks that the elder women can be remembered for are tasks such as:

• A reminder to water the plants
• A reminder to take in medicine
• A reminder to have a meal

Requirements of the secondary users:

• The secondary users need their elder woman to be healthy
• The secondary users need their elder woman to be helped to remember basic tasks

Requirements of the tertiary users:

• The tertiary users need a system that is easy to install and easy to maintain.
• The tertiary users need a system that gives a notification in case of an error or when it needs maintenance.

Approach, milestones and deliverables

Having something to care for, keeps people busy and makes life more enjoyable. For example, having a plant and hence responsibility to take care of that plant can serve as occupational therapy and ease the mind. Imagine that this plant is no ordinary plant but a robotic plant, which can signal the owner. This robotic plant can give forgetful elder women something they can nurture, but it can also give the owner reminders of basic tasks. This robotic plant could, for example, remind the owner that the plant needs water to stimulate the nurturing aspect and compliment the owner when he or she does a great job caring for the plant. The plant can also give the owner a reminder at a specific time of the day in which the plant asks if the owner has already taken their medicine or to ask if the owner has already eaten. the owner can then respond by pressing a button. In response to pressing the button, the plant can again complement the owner or stimulate the owner to do their daily activity and so the owner will create a daily routine. The plant is meant as a playful device from which elderly will benefit, namely, increase their well-being and which they will also enjoy.

Our goal is to deliver a prototype of a robotic plant that will be tested with the involvement of the users and is capable of the following actions:

• The robotic plant notifies the user that the plant needs care
• The robotic plant gives positive signals when it is properly nurtured, or stimulates the user to nurture when it is not cared for
• The robotic plant can remind the owner of small specific tasks
• The robotic plant is able to interact with the owner

With the made robotic plant our goal is to test the hypothesis, which is why it is important to finish the prototype in time.

Other deliverables are:

• A presentation which will be given in the last week of this quartile. In this presentation, we will present our prototype of the robotic plant, discuss our findings and show our progress.
• A finished wiki page which will include all our findings and progress made during this project, which will be extensively discussed.

Other milestones and deliverables can be seen in the planning below. This planning also shows a detailed task division.

About some of the topic that is discussed before, only a little research has been done as can be seen later in the literature study. Therefore some things will be tested using hypothesis tests. A hypothesis will be stated and this hypothesis will then be tested for example by taking the robotic plant to nursing homes and then conducting some questionnaires. This will be explained further in the following section called ‘Research’.

Research

What are our hypotheses?

The first hypothesis is that the user will be more interactive with a robot which has the shape of a plant than a robot shaped like a box. In the literature part, the effect of the caring character of PARO, the robotic seal, is that people with dementia started talking to the robot. A box does not have a caring character, but a plant does. So taking a robot shaped as a plant would probably be prefered over a box-shaped robot.

The second hypothesis is that a daily routine would have the same effect on men as it will have on women. As F.M. Ludwig had shown is that a daily routine will have a positive impact on women. However, this test never was executed on men, thus our second hypothesis is that this effect will be similar.

Test plan

This project is about testing two things. Namely, whether the elderly get an improved quality of life when using the robot, and secondly whether the shape of the robot plays a role in how the patients interact/react to it.

General setup

Short-term test plan

This project only has 8 weeks, in which the prototype also had to be built. So, in this project, only a short-term test plan can be executed. In the short term test, the true effect on the user’s well-being cannot be tested since this will only be visible in the long term. So, in the short-term test plan, only the impressions and opinions of the user can be tested. However, a long-term test plan is written to test whether the robot will actually improve the user’s well-being.

There is two kind of shapes that we are wanting to test here, namely the shape of a plant and the shape of a cube. Since the user has to get used to the plant, the sequence of testing will have an effect on how the user feels about the prototype. Hence there should be two groups of test-users. Group A will first test with the plant-shaped robot and then with the cube-shaped robot. Group B will first test the cube-shaped robot and then the plant-shaped robot. Both groups will have the same number of test-users. All the test users are women above the 70 in an early stage of dementia, who are living in an elderly home. Since the users have a tendency to forget things, since they have dementia, the results of a survey can be influenced by forgetfulness. To get a better view of what the user feels about the robot, a researcher will be present to analyze how the test-user interacts with the robot. The researcher also has to be present because of safety reasons. For example, the robot has a malfunction and asks the user to take their medicine every 5 minutes, if the user is very forgetful and takes medication every 5 minutes, the user can get an overdose and die. If the researcher is present, he/she can interfere and keep the test-user safe. Because of this reason, there was chosen for users that are only in the beginning stages of dementia. Users that are in later stages of dementia do not have access to their medicine without a nurse being present. A nurse would bring the user the medicine when the user has to take them in, order to guarantee their safety. Therefore it would not be possible for users in later stages of dementia to use this robot.

After having the robot in the test-users chamber for a few days, the researcher will ask the user questions in form of a survey. The survey used is attached in the Appendix. The questions for the plant-shaped robot and the cube-shaped robot are the same. In both the test with the different shapes, the robot will be placed in the same place and the same researcher will be present to analyze the user-robot interaction and to keep the user safe. The test-user may not change the robot, only interact with it. So, the user may, for example, not place pictures of his/her grandchildren on the side of the cube.

The cube and the plant are tested up front whether the microphone sounds the same in de tulip as in the cube and has the same range and volume. Moreover, the visibility of the light is tested up front, such that the lighting is approximately the same. Besides, both robots will give the same signals at the same times and both will have the same real-life plants on the side of it. So, the only thing that differs from the two robots is the shape.

Long term test plan

The effect of the plant and the impact of its shape can be tested and analysed by actually using the robot. Before the elderly will be in touch with the robot, the state of their well-being and health will be determined, using similar methods as used by F.M. Ludwig. This means that first some conversations will be held with the patients in which not only direct questions will be asked to gather knowledge and understanding of the situation, but also long conversations will be held. During the longer conversations, one can get a better impression of how the patient feels, its capability of focussing on one conversation for an extended amount of time and how the patient feels about its own well-being. Beside this conversation, a questionnaire will be created for the patient to fill in every month. Such a questionnaire should give us all the necessary information about the patient's health. Over the period in which the robot is tested, every month another conversation will be held and the patient can fill in the questionnaire again. Based on the outcomes of our own findings from the conversations and the results of the questionnaire, changes in the patient's well-being can be found.

The second test is more of a short term test plan. This will be tested by again putting the robot in the life of the user. In a few weeks the user should grow accustomed to it and can give its opinion about impressions. Such impressions are asked in the form of the survey previously mentioned. To get the best result out of the survey the questions will be open and the user can grade them from 0 to 10.

Prototype

A prototype of a robotic plant will be made in order to test our hypotheses. As for the plant a tulip was chosen and a specific system has been designed, which allows real plants to be watered if someone waters the robotic plant. Two real plants will be placed next to the robotic plant. A design of our prototype can be seen in the figure below in which the watering system will be explained and the plant idea will be shown.

Flower box

The flower box consists of two layers. The top layer consists of three parts, the two outer parts contain real plants, the middle part contains the robotic tulip. The tulip will be made completely waterproof and dirt will be placed around it. The user can water the tulip and the water will go through the dirt and into the second lower layer of the flower box. The lower layer of the flower box serves as a water reservoir. The two real plants will have a rope that connects them to the bottom layer. That rope touches the water in the lower layer and these real plants can use the rope to drink water when they need it. These plants do so by absorbing water from the rope, which the rope has absorbed from the lower layer of the flower box.

The following sections will explain the robotic plant and all of its electronic parts.

Fig. 1: Top side flower

Fig. 2: Under side flower

Fig. 4: Schematic LEDs

Fig. 5: Test LEDs

Fig. 3: Stem

3D-model

The original tulip-model came from thingiverse: https://www.thingiverse.com/thing:1429882. The tulip is made from three elements, the stem, a leaf and the flower itself. The stem and the flower were edited in FreeCAD to make room for the electronics. The inside of the flower was hollowed out to be able to install a speaker and a LED-strip (see figure 1). Four holes are extended even further into the flower to make room for four 5 mm RGB LEDs. A hole was made through the stem of the tulip all the way to the electronics to be able to wire everything (See figure 2 and 3). The rest of the electronics will be installed in the plant tray below the tulip.

The model was printed with transparent plastic. This way the LEDs can make the tulip change colours.

Electronics

There are multiple circuits inside the tulip. Every single one corresponds to a different functionality to the tulip. All circuits are driven by an Arduino UNO.

LED-strip

For this project a common anode RGB LED-strip is used inside the flower. A schematic overview of the way the LED-strip is connected is depicted in figure 4. The LED-strip is modelled by four male headers (J1 in the schematic), one corresponding to the 12V input (pin 1 in the schematic) and the other three corresponding to each colour. Eight 1.5V batteries power the Arduino and are directly connected to the LED-strip. The maximum current a colour in a single LED on the strip draws is 30mA. An I/O-pin of an Arduino delivers a maximum of 40mA which means it can only power one LED on the strip. Transistors are placed after every colour on the LED-strip to allow a bigger current to flow. Since the LED-strip has a common anode, we used NPN transistors. Thus current flows from the batteries through the LED-strip, back through all three colours and a corresponding transistor. We used BD139 transistors (view the datasheet here: http://www.redrok.com/NPN_BD135_45V_1.5A_12.5W_Hfe40_TO-126.pdf) and they can handle a maximum of 1.5A. This means enough current flows to power 50 LEDs.

The base of every transistor is connected to a PWM pin on the Arduino. The three PWM signals control the colour of every LED. Three resistors (220Ω) are placed in front of each base to limit the current.

Remaining LEDS

Four RGB LEDs are added in the bottom of the flower. They are also depicted in figure 4. In contrast to the LED-strip, these LEDs have a common cathode. They are directly connected to the Arduino, no external power source is used. Resistors are placed in front of each colour to make sure the desired current will flower. The following datasheet is used as a reference: https://www.arduino.cc/documents/datasheets/LEDRGB-L-154A4SURK.pdf. The voltage drop over the green and blue colour is 3.3V, while the voltage drop over the red colour will be 1.95V. The I/O-pin of an Arduino delivers 5V. This means there needs to be a voltage drop of 1.8V over the resistors before the green and blue colour. There needs to be a voltage drop of 3.05V over the resistor before the red colour. The blue and red colour need 30mA while the green colour needs 25mA, which means the value of the resistors will be:

[math]\displaystyle{ R = \frac{V}{I} = \frac{1.8V}{30\times10^{-3}A} = 60\Omega }[/math]

[math]\displaystyle{ R = \frac{V}{I} = \frac{1.8V}{25\times10^{-3}A} = 72\Omega }[/math]

[math]\displaystyle{ R = \frac{V}{I} = \frac{3.05V}{30\times10^{-3}A} = 101.67\Omega }[/math]

In the E12 resistor series, the closest value is 100Ω. These are also the values we used.

Water sensor

To measure if it is time for the user to give the real-life plant water a humidity and water level sensor is used. This sensor is placed in the container of the lower part of our design. Using this sensor it can be ensured that when necessary the user is asked to water the plant. Since the plant even asks the user to water the device when the water of the lower containment is not yet fully used this part of the design is made especially for when people forget to water the plant multiple times in a row. The device is easily connected to an Arduino. With the Arduino, an analogue value is available from the sensor when the board connects to water. This analogue value can thus be used to determine whether or not the plant needs water.

Fig. 5: Water sensor VMA303

Time module

Since our robot is focused on creating a good daily rhythm for the elder women it is necessary that the robot can also keep track of time. For this, we use an RTC DS1307 with AT24C32, a commonly used I2C module. The real-time clock is able to keep track of time even when the Arduino to which it is coupled is not powered. This can be done using the backup battery which can be connected. This is especially nice to have in our robot if it is connected to a wall socket to power the device. In case of a power failure, the robot won’t forget what the time is and it resumes the normal routine when the rest of the robot is powered again. The device works with an Arduino and needs to be connected to 5V, ground and two analogue pins of the Arduino to function properly. The programming can easily be done using the widely available time library written to be used to for these kinds of modules. For the first time, the module is used it is necessary to set the current time, for this many online fully written codes are available.

Fig. 5: Real Time Clock module

Speaker and audio file storage

To let the plant speak a speaker and SD card are necessary. Since the plant is quite small in the top where the speaker is meant to be placed a miniature speaker is used. This speaker can produce an output sound pressure level of 89 dB which is equal to the sound pressure of a garbage disposal or dishwasher and should be enough to ensure that the user hears it but it can’t harm the users hearing. The frequency range of the miniature speaker is 0 to 7 kHz, in which the range of the human voice fits easily. Since the Arduino does not have enough storage capabilities for the audio files, which we want to let the robot play, we need to use extra storage. For this, we use an sd-card and an external SD-card reader. We can easily record the necessary sentences using a laptop and microphone. If the product would be mass produced the quality of the recordings can be greatly improved.

Final version of the prototype

(More explaination still has to be added here)

The prototype can be seen in the figures below:

Prototype of the robotic plant

Prototype of robotic cube

Prototype of robotic plant with kitchen gardens

Testing

Testing whether or not a robot contributes to having a routine costs time and this project only consists of 8 weeks. Therefore we have set up a long term test plan, in order to test routine. However, in the short amount of time that we do have, we still wanted to test our robot. Would our plant have an effect on emotional health? How does the shape of the robot affect the elderly? And maybe some statements about creating a routine could be made if the robot would be tested for a few days. Two different tests were conducted as described earlier. These tests are a questionnaire and an interview. More details about these tests and the results will be presented in this section.

Questionnaire

A questionnaire was filled in by multiple people with several questions about how they think the robot would affect the lives of the elderly. This questionnaire can be seen in the appendix. The target audience for this questionnaire were elderly and nursing staff. For the questionnaire we travelled to a retirement home and explained to several nurses and elderly people about the robot. A short demonstration was given and features of the robot that did not work yet were explained. Then the questionnaires were given and results of this questionnaire will be given in the next section named results.

Interview

Also a more extensive test was conducted on one test person. Because of time limitations we did not have an opportunity to test the robot thoroughly on multiple people. This test was conducted in a time span of three days. Every day the robot was tested for a few hours. After this testing period an interview was held with the test person and the results will be presented in the next section.

The test person will remain anonymous for privacy reasons. However some information about this person that is relevant for this research will be presented here. The person is an elderly female and is 87 years old. She lives in a retirement home but has lived independently for a long time before she decided to move to the retirement home. She suffers from vascular dementia which is a light form of dementia and is in the early stages of this disease. Vascular dementia is one of the most common forms of dementia and is caused by brain damage from impaired blood flow to the brain. Symptoms common to this disease and the symptoms that the test person suffers from are: trouble paying attention and concentrating, reduced ability to organize thoughts or actions, problems with memory, apathy and a decline in ability to analyse a situation. These symptoms were clearly present in several moments of the given interview. For example, before the interview when there was explained what the robot does and what we are testing exactly the test person had trouble concentrating and was easily distracted by things that happened outside. Later in the interview she replied to one of the questions with a simple ‘yes’ answer and confessed that she had trouble with thinking about more extensive explanation about the question.

The robotic plant was tested in a time span of three days. The robot was tested every day for roughly for four hour per day. This time interval was chosen because the test person has to take medicine twice in this time interval. During the testing period there was asked how exactly taking the medicine worked in her retirement home. She told us that she had to take medicine three times a day: at 08:00 AM in the morning, at 17:00 AM in the afternoon and at 21:00 in the evening. At these times she has to take several pills per time as she suffers from different health problems. A specific system was created to make it easy for her to take the medicine. There was a medicine roll in her cabinet located in her own room. The pill that she has to take first was on the end side of the roll. This can be seen in the pictures below:

Figure Medicine roll

Figure Medicine roll 2

Figure Medicine information

The name of the test person has been made black, for privacy reasons. The nurse helps her take the pills, however she is also allowed to take the pills on her own. She also told us that when she was still living at home, she also took the medicine herself and had to leave the pill wrap on the table. So when the nurse visited her house, the nurse could check that she had really taken her medicine.

During the testing period we discovered that the plants that were put next to the robot did not need water often. Therefore during the testing period these plants were replaced by little kitchen gardens that were bought by a local super market. For a longer testing period the original plants would be fine, however using the kitchen gardens made the robotic plant to alert that it needed water more often. The two different types of plants were shown in the previous section about the final prototype of the robotic plant.

Results

Test person with robotic plant

Test person with robotic cube

Implementation in real life and expectations

How should this project be implemented in real life?

In the following part, the effect of the robotic plant in real life is discussed. The statements that are made are our personal expectations and ideas of the robotic plant in real life.

How would primary users interact with the device?

The primary users are women above 70 with early dementia who live at a nursing home. The robotic plant will be set somewhere in their room close to a window, where there is sunlight since the living plants next to the robotic plant should get sunlight to grow.

During the usual day of the user, the robotic plant will ask the user to give it water and possibly remind the user of their medication. The robot plant will do this via a microphone. It will then politely ask whether the primary user will water him. After the user gives water to the plant, the plant will give a positive feedback in the form of lighting. For the medication reminders, the robotic plant will ask the user whether the user has already taken her medication. This will be asked at the exact time that the medication should have been taken.

The interaction of the user with the plant will be heavily dependent on how the user feels about the plant. In many cases, the user will find it hard to accept a 'talking plant' and possibly reject it. However, the majority will accept the plant into their home. In case the user will have no intention of accepting the plant, it will be of little use. During the primary interview, questions will be asked to the user about how it feels toward such a 'strange' object to prevent a waste of time.

How would secondary users interact with the device?

The secondary users are the family of the primary user and the people working at the nursing home of the primary user and doctors and the primary user’s general practitioner. These primary users both interact the same with the device. Because of privacy reasons, the secondary users should be the ones who install the medication reminders. Information about medication is confidential and only the secondary users and the primary users know about this information. There should be a way possible to set and update these reminders. This can be done via a website or an app, which means that it can be updated from far away. This is necessary since the secondary users are not always there to implement new medication reminders or know about new medication in time to set a timer. The primary user will get new medication prescribed to them by a doctor at the hospital or by their general practitioner. These people would then be the best fit to keep track of their medication reminders. The people who work in the nursing home normally should check whether medication is taken, but now the robotic plant does this for them. If the patient did not take their medicine, an alarm should be given to the nurses, such that they can still give the medicine.

How would tertiary users interact with the device?

The tertiary users are the technicians. These users only interact with the device when there is a need for maintenance or failure.

Expectations

In the following section the expected outcomes and effects will be discussed.

Expected effect on primary users

Our expectations of the effect that the robotic plant has on primary users are the following:

Using the robot will take a significant amount of time to be useful. However, towards the end of the testing period, the user will have a more structured day due to the routine that the robot provides. This results in the user not forgetting to do certain activities, such as taking medications or eating at regular time instances, and it results in the user having a sense of time. When the user hears the second 'alarm', he/she becomes more aware of what activities could have been done at that time and thus will be more aware of its productivity and planning. In the ideal case of an actual improvement of productivity, the user will have more time to relax, play or be active, which is an essential element for maintaining health .

Another scenario is that the plant will stimulate the primary user to care for the plant, as without care the plants next to the robotic plant will die. Such a feeling of responsibility usually means that the user's mind still is very clear and the feeling of responsibility will stimulate that even more.

Expected effect on secondary users

The family of the primary user will be less worried about their elder since the robotic plant will make sure that medication is taken and the wellbeing of their elder increases.

The people working in the nursing home will have less work to do since they do not have to make sure that medication is taken by the elder, which would give them more time to do other things that are needed for a longer time period. Of course, they still have to check whether the elderly have really taken their medicine, but if the elder woman has already done this it would reduce the people of the nursing home some time. Hence this will increase the quality of the nursing home. Moreover, they will know when the elder did not take his/her medicine and hence they know when to step in and help the elder to take his/her medication. The doctors and the general practitioners will be given a little more work, however, this can be automated in the action of writing prescriptions. This would give more time to people working in the medical field then it would cost. Since this all helps the nursing homes and the elderly, this would be beneficial for the government.

Expected effect on tertiary users

The introduction of the robot will have little impact on the tertiary users

Possible extensions of the robot

As there is only limited time for this project, not all our wishes can come true. In this section future extensions, improvements and ideas of the robotic plant will be presented.

Colour testing

The robotic plant designed in this project has been used to test the hypothesis whether the shape of the plant impacts the user. This was tested by comparing a plant with a square box. Another hypothesis that is interesting to test in the future is whether or not colour signals will impact the user. For example, the plant gives green coloured lights when it is nurtured properly. It will use other colours to send reminders or to attend the user that it is not cared for.

Hearing problems

It can occur that an elderly person does not hear well and maybe this person is too stubborn to use a hearing aid. This causes the elderly person to not hear the plant properly when it gives reminders. A solution would be to install the option to higher the volume when the user does not hear the plant. The elderly person could press a button when the message has been heard. If the plant does not get a response it could repeat its message, but now with a higher volume. This cycle will be repeated until it gets a response from the user or until it reaches a certain volume level. When a certain volume level is reached it will not repeat the message anymore but will send a notification to the caring staff. However, one can object that this might become annoying to the user and causes the elderly to ignore the plant completely.

Size of the plant

Another interesting aspect to test in the future is whether the size of the robotic plant has an impact on the user. Currently, the robotic plant is a tulip and is relatively small. Maybe if it would be bigger it would have a different reaction from the user.

Kind of plant

Another aspect which would be interesting to investigate and test is whether the kind of plant matters. For this project, a tulip was used, which is a flower. However, multiple flowers exist, for example, roses or dahlias. Moreover, there exist more plants than just flowers, e.g. a tree or a cactus. Maybe if the robotic plant would look like the plants on the side, it would have a better reaction from the user. Or maybe if the robotic plant would look like the user's favourite plant, the user would like it better.

The voice of the plant

The voice of the plant is now the voice of someone on the project team, while this can feel unfamiliar to the user. Hence the impact of a familiar voice as the voice of the user's child or users wive/husband would be interesting to investigate.

Tracking system

One of the limitations of the robotic plant is that it is stationary. The robotic plant has not yet a way of knowing whether the user is present in the same room as the plant or not. So this should be fixed for later use of the plant. This can be done with a tracking device on the user. In this way, the plant could keep track of the user's location. Another example would be a wearing device which has a certain range of sending signals. So the plant will know when the user is close enough.

Voice recognition and response

It would be interesting to test whether voice recognition with a response would give a better interaction with the robotic plant.

Smart home environment

Many projects nowadays are about making smart homes to help the elderly. This robotic plant would be a wonderful addition to a smart home.

Literature Study

In order to obtain more insight into the topic a literature study about the state of the art is conducted.

General literature study about dementia therapy

Ageing and exercise: building body capital in old age.

In this article the effect of participating in group exercise is analysed. The test group, 70 year old women, corresponds to our users. The effect of being involved in group exercise is that despite their age, the women maintained bodily health and vitality, making them better able to live independently and improving their overall health. Thus being involved in group activity contributes to solving the initial problem by making the women more indepent. With our robot we expect people to have more time remaining and this often means more time for practicing group activities.

Horticultural therapy in dementia care: a literature review.

This article about horticultural therapy in dementia care combines 15 research articles together and gives a review about their findings. These research articles conducted studies about how dementia therapy affects the quality of life. Horticultural means the art and science of growing plants. The article has several findings about emotional health, self-identity and engagement. . Findings of this article are for example that gardening in groups has a positive effect as the elderly can socialize and talk about the gardening. Elderly suffering from dementia often have anxiety, agitation and depression symptoms that can lead to depression. This can have a significant effect on the quality of life. Several of these 15 studies have shown that horticultural therapy has a positive effect on these health symptoms. Since not all studies support this statement, the therapy does not support guaranteed sustained wellbeing or reducing the distress for all people. However, it represents a means by which carers can encourage elderly living with dementia to engage in meaningful activity. The main conclusion of the article is that horticultural therapy can benefit elderly suffering from dementia in several ways.

Robots that help elderly with cognitive problems

Paro

Apathy, agitation, loneliness and depression are common behavioural and psychological symptoms of dementia. These symptoms can make life distressing for the person with dementia and can also make it challenging for care staff to meet the needs of the person. In recent years, social robots have been used as a means to reduce this. Both animalistic robots, such as Paro, or robotic toys make the patients enjoy their time more and thus improve their well-being. Such an effect can also be created using plants or other systems the elder has to take care of.

Robots who take care of plants

The Pet Plant: Developing an inanimate emotionally interactive tool for the Elderly

The effect of pets on the health of elderly is widely known and this effect can also be summoned by robot pets. Another alternative is to use

Low cost colour sensors for monitoring plant growth

A relatively new area of research is that of non-destructive methods to measure the health status of plants. Such as looking at subtle changes in the colour of the leaves. This paper mainly focuses on low cost colour sensor for monitoring leaf colour of plant tissues.

To implement this system a autonomous robotic arm containing RGB colour, environmental and proximity sensors are used, as well as a camera. The robotic arm uses five stepper motors controlled through a motor controller and a micro-step driver. Their system was either compatible with the ColorPAI or the TCS 3200 colour sensor and both were controlled with a Basic Stamp microcontroller. The ColorPai uses a RGB LED light source and records the quantity of light reflected back from the object to determine the color. The TCS3200 illuminates the object with two white LEDs and interprets the colour of the object by producing a square wave with a frequency proportional to the light reflected by the object. A completely different way of determing plant leaf colour is by using an image captured by the camera and determining through software the colour of each pixel. This may however lead to false colour. For example when shadows from overlapping leaves are misinterpreted as colours of the leaves itself.

They only tested the TCS3200 and calibrated it across a broad range of green and yellow colours. In the end it had an average error around the 4% when determining an RGB color in the range of green and yellow colours.

Effect of gardens on the mood of the elderly heart

A research done on the impact of different garden design on the mood and functioning of an elderly's heart. Organised and structured gardens benefit the health of the elderly. This means that a structured placement of plants throughout a house or elderly home will also improve the health of the patient.

Garden greenery and the health of older people in residential care facilities

In this article, the relation between greenery gardens in elderly homes and the self-perceived health of the patients is examined. This was done by a vast amount of questionnaires for the residents of many elderly homes. Also elderly were asked to report on their personal health and garden visits. Both tests showed a positive and strong relationship between patient well-being and the accessibility of a garden. Visiting such a garden gives a sense of being away and even wondering about them gives a positive impact on the patients health. This information shows the importance of accessible greenery in elderly homes

Plant growth monitoring system, with dynamic user-interface

The paper develops a prototype for an efficient plant growth monitoring system. It provides data about the environmental parameters surrounding the plant and maps the changes in plant growth to the inputs given by person caring for the plants, such as the quantity of water or fertilizers. It can also measure the plant's height.

The system consists of a Raspberry Pi which analysis the output of multiple sensors and sends this data via Bluetooth to someone's phone. They also developed a user-friendly app which displays all this data. The system does not take care of the plants itself, it only monitors a variety of environmental parameters

Robotics in protected cultivation

Right now there is a lot of research to increase automation in protected cultivation. Production of high-value plants is facing an increase in problems such as increased costs for employees and decrease in skills. Robotics can help decrease these problems. The problems can be sorted into different groups. For seeding, grafting and cutting there are already products which can do this just as there are for transporting, sorting, packaging and cleaning. However, there is still a lot to win in the department of weeding, thinning, leaf picking, protection and harvesting. The technical challenges in robotics are the fact that technologies have to deal with complex environments. There are two solutions which can be used in this case, advancing technology or modifying the environment.

Thermostats and measurements of oxygen(state of the house)

NeoFox Oxygen:

This oxygen sensor shows oxygen sensing using two methods. First of all, by making an electrochemical compound that will conduct current based on oxygen levels in the material. This is a relatively easy method and gives linear dependency between the oxygen level and the current created, but will also consume the oxygen. Secondly, by using a fluorescent material that will react to oxygen, different sorts of light will be radiated. Based on the frequency, amplitude, and phase of this emission, the amount of oxygen can be determined. This second method may be faster and more reliable, but also more difficult to produce.

PhO2: Smartphone-based Blood Oxygen Level Measurement Systems using Near-IR and RED Wave-guided Light

Measuring a patient's blood oxygen level plays a critical role in healthcare practices. In the paper, they develop a phone-based oxygen level estimation system which uses the camera and flashlight functions of a smartphone. Since the camera and the flashlight of a smartphone aren't made for this purpose, they encountered many challenges in using them for this purpose.

Blood oxygen level is often indicated by oxygen saturation measurement (SpO2). Accurately measuring SpO2 with a high frequency is critical in monitoring the well being of key organs and also to provide warning signs of abnormalities. One of the most common ways to measure SpO2 is by using a pulse oximetry system. This system comprises dedicated hardware and software. It works by project light beams at specific wavelengths deep into the users' finger, toe, earlobe, or other location. This light hits dedicated photoelectrodes and the intensity of the light received carries information that can be used to determine the users' SpO2. This way of measuring SpO2 is very reliable but it requires the user to purchase one and to carry it with them during the day. Patients often forget to take the device with them or forget to charge it. Also, different patients have very different finger and earlobe sizes and the devices are thus not always a good fit.

There are already many applications for smartphone out there that use the flashlight and camera to determine various blood properties. However, none have the ability to accurately estimate SpO2. This low accuracy is a result of the fundamental challenges when one tries to repurpose the camera and flashlight for SpO2 measurements. Examples of these challenges are:

• One needs an IR wavelength
• One needs to work around the varying movement, pressure, contacting area of the users' finger

In the paper, they develop a hardware add-on (PhO2) designed to be snapped on the phone as a phone case which has optical filters of different wavelengths. The add-on helps stabilize the users' finger. Due to the limitations of the phone's hardware, the reflected light captured by the PhO2 is further processed by dedicated algorithms.

Smart device for gas range

This is a smart device, which knows when the gas is on and whether the stove is left unattended. In case of an unattended stove, the device will send a message to the user, which can shut off the gas supply or call for relevant party's immediate attention remotely. When sending the message fails or when there is no response, the device will automatically shut off the gas supply.

Iot-Based Intelligent Modeling of Smart Home Environment for Fire Prevention and Safety

Fire detection has become an issue since it caused severe damage including the loss of human lives. To reduce property damage and save lives, early detection of a fire event is very effective. The installation of a fire alarm system is the most convenient way to detect a fire early and avoid losses.

In the paper, they propose an efficient, IoT-based intelligent home fire prevention system using multiple sensors, which each uses its own mechanism for detection.

In the paper, problems and challenges related to the current approaches are identified, GSM communication is used to alert the user at early stages, star topology is used for the deployment of sensors and communication between sensors and main home sink and the system concerning energy consumption is evaluated.

Design and Development of a Low-Cost, Portable Monitoring Device for Indoor Environment Quality

In this article, the design and development of a low-cost, portable monitoring system for indoor environment quality(IEQ) is described. A prototype is made with commercially available low-cost sensors and a do-it-yourself approach is provided. The designed system monitors temperature, humidity, PM2.5, PM10, TVOC(x3), CO2, CO, IAQ, illuminance, and sound levels. The biggest advantage of this design is the low cost, since it provides a comprehensive, portable, and real-time monitoring solution, for less than 200 dollars.

A cheap and third-age-friendly home device for monitoring indoor air quality

This article proposes a new methodology to analyze indoor air quality with a cheap and third-age-dedicated device. The researchers developed a prototype, which they called HOPES, Home Pollution Embedded System. This prototype gives simple and understandable information, also comprehensible for people with cognitive problems or that are not familiar with new technologies. The prototype gathers data about pollutants and displays the different air pollutants concentrations to the user. These air pollutants concentrations are from toxics gasses up to explosives. Furthermore, in the paper, an overall air quality index is elaborated and displayed by HOPES with lights and numerical information. HOPES is an internet of things device. Hopes works in real time and can be connected to a geographic information system platform and the web to add spatial information about each pollutant. The results highlighted how it is possible to get useful air quality information with a cheap device.

Medication reminders

Smart Home medication reminder system

Many elderly people need assistance from nurses, housekeeping, and visitors that remind them to take their pills. New technologies can provide a medication reminder system. This article mainly focuses on the Open Home Automation Bus (openHAB). This is a software that integrates and thus combines different home technologies together. There are several smart technologies that help elderly people and this software can combine them and make it easy to use.

A WSN smart medication system

It is very important that patients or elderly people take in their medicine correctly. However, for elderly, it often becomes difficult to remember whether or not they have already taken their medicine. Also when these people have lots of different pills it can become hard to remember which to take and when to take them. A wireless sensing network (WSN) system have been invented which can remind patients to take their medicine. The system consists of a master panel (MP) and portable smart pill-boxes (SPB). Magnetic sensors are used to detect whether or not there are pills in the (SPB). If not pill has been taken. This paper describes the three-state pill sensing mechanism and all other aspects of the wireless sensing network.

Medicine Reminder and Monitoring System for Secure Health Using IOT

Elderly people sometimes tend to forget doing basic things among daily routine. This can lead to them forgetting to take their medicine at the right time of forgetting it entirely. Using the Internet of Things (IoT) network low-cost medical sensing can be produced. There have been multiple tests with technologies to find a way to decrease this problem. A monitoring system and sensor can send information using a wireless module. The information can be shared using IoT, however, since health information can be very personal an encryption or decryption purpose is recommended.

MESSAGE INA BOTTLE

Medication reminders in the form of pill bottles are already available. The bottle can notify patients when they need to take their medication or missed a dose as well as seeing that the pills are almost out and notifying a pharmacy. This technology can not only increase the effectiveness of certain drugs but also reduce the readmission rates at hospitals. The pill bottle uses a sensor to detect opening and closing of the lit and compare its content. This is then sent to the startup's server which can analyze the data. Using different colours of light the user is given an indication of when taking medication is necessary.

Robots who look at the eating of their users

An Intelligent Food-Intake Monitoring System Using Wearable Sensors.

Researchers are looking for accurate methods requiring less user-involvement to assess general food-intake. The paper proposes an intelligent foot-intake monitoring system that can automatically detect eating activities. An in-ear microphone with a miniature camera is combined in a light-weight wearable headset. The sound from the microphone is classified into different eating activities and the camera takes pictures of the food if a chewing activity is detected. The key images of food are then selected sequentially and a dietary assessment log is generated to reflect a user's dietary behaviour. Novelties in this design were:

• Developing a noise-resilient sound activity detection method suitable for daily use
• Introducing food images to improve assessment accuracy
• Selecting key images automatically to minimize the size of the food-intake assessment log

Big fridge is watching you

The smart kitchen is becoming reality. With the enormous increase of technology in everyday things one cannot deny the probability of a smart fridge. This fridge could use information from your wearables, agenda and smart devices to recommend meals. The used data can also be relayed to doctors or caregivers, for instance when an elderly is not eating enough. Furthermore, it can help people who tend to forget what is in the fridge with keeping an inventory and tracking expiration dates. This way an order could be placed for what is needed from the supermarket.

Automatic Dietary Monitoring Using Wearable Accessories

In this article, a lot of research is done on how they modeled a person dietary. However, this is not really relevant to our project. But, later in the article a discussion is given about the state of the art of sensing technologies, integration in accessory-based wearable devices and estimated parameters of different dietary dimensions. After which the researchers explain which challenges must be addressed to make ADM technology viable.

The Billy-Billy robot

Fig. 1: Billy-Billy robot

At the end of our literature study a robot was found that was made using similar ideas as our project.

The robot that we discovered is called the Billy-Billy robot. It is a robot in the shape of a face with a plant growing out of its head. Billy-Billy is a flowerpot that has several unique features. It is an interactive flowerpot who makes the life of elderly people more enjoyable in an interactive way. Billy-Billy has smart sensors that can detect when the herbs or plant growing out of Billy-Billy needs more light or extra water. It can then notify the elderly. Billy-Billy is easy to install and can be used in several minutes and also has plant seeds included when bought. It does require a network connection. This internet connection provides an online platform in which family and friends can see how happy Billy-Billy is. It also allows the family and friends to send text messages to Billy-Billy and in return Billy-Billy can communicate these messages to the elderly. Lastly Billy-Billy has an integrated agenda system which notifies the user to take medicine, notify when a carer arrives or when a family party is planned.

We are very interested in the Billy-Billy robot and contacted the company to ask specific questions about their robot. This robot shows that we as a group are not the only one interested in the subject and that this kind of robot really could help elderly people.

Appendix

Survey

The survey used for testing the differences between shapes will feature the following questions:

• How long has the robot been implemented to your home?
• What is your general impression of the robot?
• Does the robot give you a sense of time based on its alarms?
• Do you think the robot is an addition to your house?
• Does the robot encourage you to take care of the plants?
• Do you feel like taking care of the plants if the robot asks you to do it?
• Do you like the appearance of the robot?
• Does the robot help you remember to do other tasks?
• Does the robot encourage to be more active?
• Do you notice being more aware of your surroundings?