Assignment 2 for Machne Learning with Robotics

1. PID controller

This assignment is about writing a PID control program to let the tribot follow a wall in a certain distance. Test different settings for the control parameters and write up your findings in form of an extended abstract. The extended abstract must fit on one letter-size page and should include a title and the names and affiliation of the investigators.

2. Localization

Assemble a simple robot arm made out of one LEGO motors with a pointer attached to the orange turning plate. This simple robot arm is driven periodically by by the controller waving.m

The position of the motor can be estimated in various ways. The odometry model is given by the function forward.ReadFromNXT.Position(). The popsition can also be estimated from the motion model. Finally, we can also estimate the position of the motor using a webcam. For this we provide two functions, one that sets up the webcam capturing parameters, provided in file setup_sensor.m, and the actual function that gives a point estimate of the position using the camera, provdided in file sensor_model.m.

The aim of this assignment is to compare the estimates from the motion model alone, the sensor model alone, and a combination of the two, to the odometry measurement. The combination of the motion model and sensor model should achieved by Bayesian Filtering.

Plot the different estimations versus odometry model and explain your findings. Explain how you applied the motion model and the sensor model to a Bayesian filtering algorithm.

Submit your answers to with subject line A2 in pdf format. Your submission has to be received by 6pm on Thursda , Oct. 6.