Towards a Viable Autonomous Driving Research Platform
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This paper presents the design and implementation of a highly integrated autonomous driving research platform developed at Carnegie Mellon University, based on a stock Cadillac SRX. The research addresses the critical gap between laboratory prototypes and commercially viable autonomous vehicles, focusing on reliability, safety, cost-effectiveness, and social acceptance. Unlike previous platforms that featured heavy exterior modifications or non-automotive-grade sensors, this platform prioritizes minimal appearance changes and the integration of production-ready components to demonstrate the feasibility of consumer-oriented autonomous driving. The vehicle’s architecture comprises several key subsystems. The drive-by-wire (DBW) actuation system enables computer control of braking, throttle, steering, and gear selection through mechanical and electrical modifications, including an electromagnetic clutch for steering that allows safe manual override. A hierarchical control system tracks trajectories with lateral errors under 30 cm. The perception system utilizes a hybrid sensor suite of six LIDARs, six radars, and eight cameras, providing 360-degree coverage with redundancy; localization combines GPS/RTK with lane-feature-based local positioning to achieve centimeter-level accuracy. To support these systems, an isolated auxiliary power system with uninterruptible power supplies ensures critical DBW functions remain operational during failures. The computing platform employs four identical mini-ITX computers with GPUs, arranged in a fault-tolerant architecture with redundant network switches to prevent single points of failure. The platform demonstrates a wide range of autonomous behaviors, including lane keeping, lane changing, intersection handling (both with and without V2I/V2V communication), and detection of pedestrians, bicyclists, and work zones. Safety is enforced through a three-layer system: remote stop, auto/manual switching that physically disengages actuators, and an emergency stop that powers down the auxiliary system. The user interface supports four levels of interaction, from mission planning to remote access, facilitating smooth human-machine cooperation. The significance of this work lies in its demonstration that a robust, safe, and intelligent autonomous vehicle can be built with minimal aesthetic compromise, addressing key barriers to social acceptance and commercial deployment. By integrating fault-tolerant computing, redundant sensing, and production-compatible hardware, the platform provides a viable foundation for future research into socially cooperative driving behaviors and extensive real-world testing.
Key finding
The integrated autonomous driving platform successfully demonstrated robust performance in lane keeping, intersection traversal, and obstacle avoidance during extensive public road and closed-field testing, maintaining safety through redundant hardware and a three-layer fail-safe system.
Methodology
field_study
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 19 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Methodological Resource: tool software
- Theoretical Contribution: computational model