Monte Carlo Localization (Particle Filter) скачать в хорошем качестве

Monte Carlo Localization (Particle Filter)

Part of a project for my Computational Basis of Intelligence class was to implement a Particle Filter, so I chose to implement the Monte Carlo Localization algorithm. The landmarks are known and unambiguous. The PF is used to track the position of the robot and provide measurement updates. In the simulation, 10000 particles are used to estimate the state. There is noise in the system and the algorithm attempts to compensate for the noise. The agent (blue robot) gives commands it wants executed (control commands like go forward, turn) and then asks if any objects have been sensed. It then tries to calculate its position using this knowledge. The red is the actual location The blue (agent) is the believed location and is the average of the particles The green is a simulated LIDAR sensor.