How nishita1652/ecological-pcc-ev works
Eco-PCC is a proof-of-concept EV cruise-control system: synthetic traffic scenarios support LSTM training, and the runtime predicts leader speeds, optimizes ego-car torque, simulates vehicle physics, and compares the result with a reactive-driver baseline. The simulation reports speed, safety-gap, and energy outcomes in a dashboard. The README documents forecasting, safety, comfort, and energy objectives; edges reflect only relationships visible in the excerpts or README. The trained weights are present, but their internal structure was not inspected.
Main entry points: run_simulation.py Race simulation · lstm_model.py LSTM training · mpc_controller.py MPC controller · data_generator.py Traffic scenarios
Components 10
Traffic Data
- Traffic scenarios
src/data_generator.py
Forecasting
- LSTM training
src/lstm_model.py - Traffic weights
models/traffic_lstm.keras
Predictive Control
- MPC controller
src/mpc_controller.py - EV dynamics
src/vehicle_dynamics.py
Simulation
- Race simulation
src/run_simulation.py - Race traffic profile
src/run_simulation.py - Reactive driver
src/run_simulation.py - Results dashboard
src/run_simulation.py
Other components
- Benchmark operator
Connections 11
- Benchmark operator → Race simulation: runs benchmark
README.md - Traffic scenarios → LSTM training: generates sequences
src/lstm_model.py - LSTM training → Traffic weights: saves weights
src/lstm_model.py - MPC controller → Traffic weights: loads model
src/mpc_controller.py - Race simulation → Race traffic profile: generates traffic
src/run_simulation.py - Race simulation → MPC controller: predicts and optimizes
src/run_simulation.py - Race simulation → EV dynamics: updates EVs
src/run_simulation.py - Race simulation → Reactive driver: computes torque
src/run_simulation.py - Race simulation → Results dashboard: plots outcomes
src/run_simulation.py - LSTM training → Traffic scenarios: imports generator
src/lstm_model.py - MPC controller → EV dynamics: uses EV state
src/mpc_controller.py