CADET: A Modular Platform for Evaluating Distributed Cooperative Autonomy in Connected Autonomous Vehicles
Pragya Sharma, Brian Wang, Mani Srivastava
AI summary
Problem
Existing autonomous vehicle simulators overlook the systems-level challenges of cooperative autonomy, such as network latency, compute heterogeneity, and multi-tenant contention, leaving a gap in systematically evaluating how distributed deployment choices impact real-world safety.
Approach
CADET disaggregates the autonomous vehicle stack into composable modules that can be flexibly deployed across vehicles, roadside units, and edge/cloud servers, paired with NetWaggle to emulate realistic V2X network dynamics and heterogeneous compute constraints.
Key results
- Introduction of CADET, a modular platform for reproducible distributed cooperative autonomy evaluation
- Development of NetWaggle, a network emulation layer capturing realistic V2X delays and device heterogeneity
- Demonstration that V2V intent packets outperform cloud-based perception under adverse network conditions
- Open-source release with benchmarks and hardware-in-the-loop validation capabilities
Why it matters
It enables systems and ML researchers to systematically benchmark distributed inference and cooperative autonomy policies under realistic constraints, accelerating the development of safe, scalable connected vehicle ecosystems.
Abstract
Deep learning models are increasingly central to autonomous vehicle (AV) pipelines, yet their integration has traditionally followed a monolithic design where perception, planning, and control execute on a single onboard computer. This design overlooks the emerging paradigm of cooperative autonomy, where vehicles interact with roadside units (RSUs), edge servers, and cloud-hosted intelligence through vehicle-to- everything (V2X) connectivity. Cooperative perception and con- trol improve safety and efficiency, but also introduce systems- level challenges: network latency, compute heterogeneity, and multi-tenant contention, all critically affect real-time decision- making. These challenges are further amplified by the increas- ing reliance on large foundation models, whose scale necessitates cloud deployment. We present CADET (Cooperative Autonomy through Dis- tributed Experimentation Toolkit), a modular platform for systematic and reproducible evaluation of distributed coopera- tive autonomy systems under realistic deployment conditions. CADET decouples the AV stack into composable modules that can be flexibly deployed across vehicles, infrastructure, and edge/cloud tiers. The framework integrates state-of-the- art models, incorporates trace-driven network and workload emulation, and provides synchronized model-, system-, and task-level instrumentation. Through V2V and V2I experiments, we show that distributed deployment choices fundamentally shape safety, with V2V intent packets outperforming cloud- based perception and RSU-assisted perception sustaining safety until overloaded by concurrent requests. Although designed for AV pipelines, CADET also supports dataset-driven experimen- tation, enabling systems and ML researchers to benchmark distributed inference workloads independently of full vehicle simulation. CADET is open source, with code and demo available at https://nesl.github.io/cadet-web.