PETcat - Obstacle Avoidance

Path Planning with Gaussian Potential Maps

The code can be found here.
To determine the optimum value for gamma, we assigned rewards for a path; depending on the number of collisions with obstacles, the number of steps required to reach the goal and whether the agent was able to reach the goal under a threshold number of steps.
Preliminary tests showed that the value for gamma should lie between 0 and 1.5.

Results :

Test 1

[Moving from (0,0) to (4,-2)]

Best Path (Gamma = 1.47) :

Test 2

[Moving from (0, 0) to (5, 5)]

Best Path (Gamma = 0.59) :

Both environments work well with Gamma = 1

With Mobile Obstacles

We realised the need to use lower values for Gamma.
Best Path : Gamma = 0.59