Retake Embedded Motion Control 2018

Follow Me!

Context

An increase of robotic deployment is nowadays observed; one could think about care robots for elderly care, AGV's for industrial application, or even a simple vacuum robot in a typical household. Repositioning these robots often occurs manually as, e.g., one picks them up or controls them (via a joystick) to a given position. This, however, may not always be desired as you may for example want the robot to carry something with you (maybe a table). We thus ask you guys to design a person following algorithm to solve this problem.

Task description

The main objective of this assignment is to follow a person in a cluttered environment - more details about the environment are given further below - for a time span of 2 minutes. The Person Of Interest (POI) will walk a pre-defined route, and on important points we will put down so-called 'target strips' which the POI will cross. To test whether PICO is indeed able to follow the POI correctly, PICO should also cross these target strips. A brief overview of the setup is shown on Figure 1.

Some specifications of interest:

• The POI will start initially close to PICO at standstill. After PICO starts following him, the POI will move with a velocity greater than 0 meter per second, but no more than 0.5 meter per second. The POI will furthermore move in such a manner that (in most cases) two legs are visible.
• The target strips can be regarded as markers on the floor which PICO can't recognize. These target strips have a width of 1 meter, and the POI will pass through the middle of these target strips. The number of target strips should be around 5.
• We recommend that PICO follows the POI at a distance of approximately 0.5 meter; this will guarantee that PICO crosses the target strips.
• A failure is stated when the following observations are made: 1. PICO is unable to cross 80% of the target strips 2. PICO bumps into the POI. If a failure is determined, then you can retry once (thus a total of 2 takes).

Figure 1: A graphical representation of the setup of the assignment. The pre-defined path is indicated by the dashed lines, the target strips by the black lines perpendicular to the pre-defined path, the POI that follows the pre-defined path in blue, and PICO in red.

Regarding the cluttered environment, three 'levels' are designed ranging from easiest to hardest. To pass this course, we ask you to at least clear level 1, i.e., the easiest task. Clearing any subsequent levels will naturally allow you to pass this course with a better grade.

Level 1 (grade 6)

In this first level we have other actors present in the area of interest. These actors will similarly as the POI start at standstill. After PICO has found his POI and has started his following-algorithm, all other actors will also walk around. The only restriction that the actors have is that they can never occlude the POI, i.e., they can not walk in between PICO and the POI. Figure 2 depicts a situation which can not occur.

Figure 2: A graphical representation of a situation which can not occur for Level 1. In red we see PICO, blue the POI, and in darkgreen the actor.

Level 2 (grade 8)

In this second level we have only static obstacles present. The difficult part w.r.t. the previous level is that a path will be defined such that occlusion will be present, i.e., the legs of the POI will dissapear behind a static object and hence complicate the following algorithm. When such an occlusion occurs, PICO is allowed to shout 'stop' to find the POI again. Note that this is limited to a time interval of 15 seconds, after which the POI will move again. When PICO has found the POI, it should likewise say 'go' to allow the POI to walk further.

Level 3 (grade 10)

In this level, both static objects and actors are present. All the difficulties of the previous levels are present with furthermore no restrictions on the actors to prevent situations as depicted on Figure 2.

Other relevant information

We will grade you on the basis of your performance