sanjay8906/klh-cse-2026-27-adas-15-rl-based_autonomous_intersection_navigation

How sanjay8906/klh-cse-2026-27-adas-15-rl-based_autonomous_intersection_navigation works

This is a research ADAS simulator for reinforcement-learning-based intersection navigation. The interactive workflow configures a scenario, runs MetaDrive, fuses sensor inputs into perception, and uses a vehicle controller to produce driving actions; a sensor display supports inspection. A separate RL workflow trains and evaluates a policy in a Gymnasium environment. The graph reflects documented capabilities and source-index references, but several implementation links are based on imports rather than sampled call sites, so uncertain connections are omitted.

Main entry points: sensor_fusion.py Sensor Fusion · adapter.py MetaDrive Adapter · main.py Simulation CLI · environment.py RL Environment

Components 21

Interactive Simulation

Simulation CLI src/main.py

Scenario and Simulator

MetaDrive Adapter src/adas_sim/metadrive/adapter.py
Scenario Configurator src/adas_sim/scenarios/scenario_configurator.py
Map Loader src/adas_sim/map/map_loader.py
Indian Traffic Env src/adas_sim/metadrive/indian_env.py
Traffic Manager src/adas_sim/metadrive/traffic_manager.py
MetaDrive

Sensors and Perception

LiDAR Sensor src/adas_sim/sensors/lidar.py
Radar Sensor src/adas_sim/sensors/radar.py
Ultrasonic Sensor src/adas_sim/sensors/ultrasonic.py
Camera Sensor src/adas_sim/sensors/camera.py
GPS Sensor src/adas_sim/sensors/gps.py
IMU Sensor src/adas_sim/sensors/imu.py
Vehicle State src/adas_sim/sensors/vehicle_state.py
Sensor Fusion src/adas_sim/sensors/sensor_fusion.py

Reinforcement Learning

RL Environment src/adas_sim/rl/environment.py
Training Pipeline src/adas_sim/rl/training.py
Training Entry scripts/train_rl.py

Control and Display

Vehicle Controller src/adas_sim/controllers/vehicle_controller.py
Sensor Display src/adas_sim/gui/sensor_display.py

Other components

User

Connections 26

User → Simulation CLI: starts README.md
Simulation CLI → MetaDrive Adapter: runs simulator src/main.py
Simulation CLI → Sensor Fusion: processes perception src/main.py
Simulation CLI → Vehicle Controller: gets actions src/main.py
Simulation CLI → Sensor Display: updates display src/main.py
MetaDrive Adapter → Scenario Configurator: configures scenario src/adas_sim/metadrive/adapter.py
MetaDrive Adapter → Map Loader: queries map src/adas_sim/metadrive/adapter.py
MetaDrive Adapter → Indian Traffic Env: uses environment src/adas_sim/metadrive/adapter.py
Indian Traffic Env → Traffic Manager: uses traffic src/adas_sim/metadrive/indian_env.py
MetaDrive Adapter → MetaDrive: configures and resets src/adas_sim/metadrive/adapter.py
MetaDrive → MetaDrive Adapter: returns observations
Sensor Fusion → LiDAR Sensor: processes scans src/adas_sim/sensors/sensor_fusion.py
Sensor Fusion → Radar Sensor: processes detections src/adas_sim/sensors/sensor_fusion.py
Sensor Fusion → Ultrasonic Sensor: processes readings src/adas_sim/sensors/sensor_fusion.py
Sensor Fusion → Camera Sensor: uses imagery src/adas_sim/sensors/sensor_fusion.py
Sensor Fusion → GPS Sensor: processes position src/adas_sim/sensors/sensor_fusion.py
Sensor Fusion → IMU Sensor: processes motion src/adas_sim/sensors/sensor_fusion.py
Sensor Fusion → Vehicle State: processes state src/adas_sim/sensors/sensor_fusion.py
Vehicle Controller → Sensor Fusion: reads perception src/adas_sim/controllers/vehicle_controller.py
Vehicle Controller → MetaDrive: controls vehicle
Simulation CLI → RL Environment: can run policy
Training Entry → Training Pipeline: starts training scripts/train_rl.py
Training Pipeline → RL Environment: trains and evaluates src/adas_sim/rl/training.py
RL Environment → MetaDrive Adapter: wraps simulator src/adas_sim/rl/environment.py
RL Environment → Sensor Fusion: builds perception src/adas_sim/rl/environment.py
RL Environment → Vehicle Controller: uses controller src/adas_sim/rl/environment.py

Drawn by GitDiagram on November 14, 2026 from the repository's file tree, README and source files. Read as Markdown ·