SmartLoader

Welcome to SmartLoader! Innovation project February-May 2020.

All code needed to train and implement imitation and reinforcement learning on simulation and real-life loader.

Includes:

gym_SmartLoader - gym environment for:
- bobcat simulation (SmartLoader_env)
- Benny's heat map bobcat simulation (HeatMap_env)
- Algoryx simulation (push_algoryx)
Sim_Agents: algorithms for RL (train_agent) and BC agents for simulation
SmartLoaderIRL - Environment for real-life scenario
Real-life episode runners: episode_runner_push_mission, episode_runner_lift_mission
recordings_manipulations - save data from real-life joystick recordings
Imitation learning from recordings - Pos_EstimatorAgent_CNN_Keras
Benny’s heatmap - HeatMapGen
LLC - PID for controlling all loader’s actions (or bobcat in simulation): steer, speed, lift, pitch. Different PIDs for different missions: push, dump, drive back, load.
Algorix training

Name		Name	Last commit message	Last commit date
Latest commit History 112 Commits
LLC		LLC
Sim_Agents		Sim_Agents
gym_SmartLoader		gym_SmartLoader
saved_experts		saved_experts
saved_models		saved_models
stable_bl/PushStonesEnv/model_dir/sac		stable_bl/PushStonesEnv/model_dir/sac
Blade_Height_Agent_SL_keras.py		Blade_Height_Agent_SL_keras.py
Blade_Height_Agnet_SL_talos.py		Blade_Height_Agnet_SL_talos.py
Conf_Agent_SL_keras.py		Conf_Agent_SL_keras.py
Conf_Agent_SL_talos.py		Conf_Agent_SL_talos.py
Env_runner.py		Env_runner.py
HeatMapGen.py		HeatMapGen.py
HeatMapGen_old.py		HeatMapGen_old.py
KERAS_conf_model_1.7z		KERAS_conf_model_1.7z
KERAS_pos_est_model_0.7z		KERAS_pos_est_model_0.7z
Pos_EstimatorAgent_CNN_keras.py		Pos_EstimatorAgent_CNN_keras.py
README.md		README.md
SmartLoaderIRL.py		SmartLoaderIRL.py
episode_runner_lift_mission.py		episode_runner_lift_mission.py
episode_runner_push_mission.py		episode_runner_push_mission.py
improved_train_agent.py		improved_train_agent.py
recordings_manipulation.py		recordings_manipulation.py

Provide feedback