search

reiniscimurs / DRL-robot-navigation

Deep Reinforcement Learning for mobile robot navigation in ROS Gazebo simulator. Using Twin Delayed Deep Deterministic Policy Gradient (TD3) neural network, a robot learns to navigate to a random goal point in a simulated environment while avoiding obstacles.
MIT License
487 stars 97 forks source link

Selection of local target points #79

Closed ZCPRL closed 3 months ago

ZCPRL commented 8 months ago

Hello, I have not found the code of the part that selects the local target point in the project, may I ask where the code of this part is implemented?

reiniscimurs commented 8 months ago

Goal point is set by calling the change_goal function: https://github.com/reiniscimurs/DRL-robot-navigation/blob/main/TD3/velodyne_env.py#L320

ZCPRL commented 8 months ago

Thank you for such a quick reply. Your team's processing of laser data is to select points with the smallest distance from each Angle gap within 180 degrees and take these points as state values. My question is whether each Angle gap point is a local target point. Another problem is that the code does not see the calculation of the evaluation (h(ci)) of each local target point. According to the content of the paper, the point with the smallest h(ci) is selected as the local target point, looking forward to your answer.

1698659126373

reiniscimurs commented 8 months ago

Hi,

My question is whether each Angle gap point is a local target point.

I do not understand this question. What do you mean by local target point? Each gaps minimal laser reading is taken as state information. It has no bearing on goal position selection as it is a random point in the environment.

I think you might be looking at the wrong repository. This repo is just for training a generic DRL-based motion policy. The exploration algorithm (where a version of this policy is used) is available in the GDAE repository. The calculation of POI is given in: https://github.com/reiniscimurs/IDLE-heuristic

ZCPRL commented 8 months ago

I see. Thank you for your answer!