MAAD Highway: Anomaly Detection in Multi-Agent Driving
What it is
This repository provides the official code, simulation environment, and dataset for the CoRL 2021 paper on anomaly detection in multi-agent driving trajectories. It implements a Spatio-Temporal Graph Auto-Encoder (STGAE) and baseline methods to detect abnormal driving behaviors in highway scenarios.
The problem it solves
Detecting subtle, multi-agent anomalies in driving trajectories that single-agent models or simple reconstruction errors often miss.
What you could use it for
- Training and evaluating anomaly detection models for autonomous driving
- Simulating multi-agent highway interactions for data generation
- Benchmarking graph neural networks for trajectory reconstruction
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.