# Harness Engineering for Physical AI: Robot Middleware Is the Harness Layer

arXiv: https://arxiv.org/abs/2606.09416 · HTML: https://mlsanity.com/paper/2606.09416/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 36 references that are in the map.
- First posted on arXiv 8 Jun 2026.

## Builds on (most cited first, 13 of 13)

- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks](https://mlsanity.com/paper/1610.02136/) (2016)
- [GR00T N1: An Open Foundation Model for Generalist Humanoid Robots](https://mlsanity.com/paper/2503.14734/) (2025)
- [Safe Reinforcement Learning via Shielding](https://mlsanity.com/paper/1708.08611/) (2017)
- [Serving DNNs like Clockwork: Performance Predictability from the Bottom Up](https://mlsanity.com/paper/2006.02464/) (2020)
- [Neural Simplex Architecture](https://mlsanity.com/paper/1908.00528/) (2019)
- [Safety Aware Task Planning via Large Language Models in Robotics](https://mlsanity.com/paper/2503.15707/) (2025)
- [Harnessing Embodied Agents: Runtime Governance for Policy-Constrained Execution](https://mlsanity.com/paper/2604.07833/) (2026)
- [Guardian: Safe GPU Sharing in Multi-Tenant Environments](https://mlsanity.com/paper/2401.09290/) (2024)
- [Optimal Resource Efficiency with Fairness in Heterogeneous GPU Clusters](https://mlsanity.com/paper/2403.18545/) (2024)
- [PAAM: A Framework for Coordinated and Priority-Driven Accelerator Management in ROS 2](https://mlsanity.com/paper/2404.06452/) (2024)
- [In Serverless, OS Scheduler Choice Costs Money: A Hybrid Scheduling Approach for Cheaper FaaS](https://mlsanity.com/paper/2411.08448/) (2024)
- [Argus: Quality-Aware High-Throughput Text-to-Image Inference Serving System](https://mlsanity.com/paper/2511.06724/) (2025)

## Built on by (most cited first, 3 of 3)

- [JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolution](https://mlsanity.com/paper/2608.25593/) (2026)
- [Ensuring Safe Physical AI in Urban Mobility via Hazard-Informed Synthesized Envelopes](https://mlsanity.com/paper/2608.14481/) (2026)
- [Task-Specified Active Metrological Inspection with Measurement-Steered VLA Manipulation and Deterministic Evidence Gating](https://mlsanity.com/paper/2609.14219/) (2026)

## Abstract (arXiv)

Robot middleware faces a new role in the era of Physical AI. Learned policies, planners, and vision-language-action (VLA) models now enter deployed robots as causal participants on the control path, but the layer that integrates them with timing, scheduling, and network has not been named. Recent language-agent work names this layer the harness, the external system that mediates tools, manages state, bounds resources, and records execution. The robotics community has not yet adopted this framing, and we propose that robot middleware is that harness. A Physical AI harness differs from a software harness in where it intervenes. A software harness mediates at tool-call boundaries. A Physical AI harness must mediate at control, computing, and communication simultaneously, because a learned policy's output crosses all three: its commands shift the trajectory, its inference time shifts the schedule, and its payload shifts the bandwidth. Robot middleware is the lowest robot-stack layer with mediating abstractions over all three, so it is best positioned to compose their enforcement. It already provides most of what a harness needs but lacks the enforcement for an AI model. We name this missing enforcement as three functions: Projection gates each output at emission, Isolation bounds the model's execution and transmission slot, and Transfer falls back to a verified baseline when checks fail. Each appears today as hand-built application code in deployed robot systems, built on surfaces robot middleware already provides. Robot middleware should host them not as the best single-axis enforcer but as the layer that composes all three. We sketch this as a ROS 2 Harness Profile, a deployment artifact that carries an AI model's declared output region, inference budget, and operating regime while the middleware enforces them across ROS 2, DDS, and Zenoh.