# RoCUS: Robot Controller Understanding via Sampling

arXiv: https://arxiv.org/abs/2012.13615 · HTML: https://mlsanity.com/paper/2012.13615/

## Key facts

- Cited by 6 papers in the map (data through 7 Oct 2026).
- First posted on arXiv 25 Dec 2020.

## Built on by (most cited first, 6 of 6)

- [Deep Surrogate Assisted Generation of Environments](https://mlsanity.com/paper/2206.04199/) (2022)
- [Surrogate Assisted Generation of Human-Robot Interaction Scenarios](https://mlsanity.com/paper/2304.13787/) (2023)
- [Enabling Adaptive Agent Training in Open-Ended Simulators by Targeting Diversity](https://mlsanity.com/paper/2411.04466/) (2024)
- [AutoPreview: A Framework for Autopilot Behavior Understanding](https://mlsanity.com/paper/2102.13034/) (2021)
- [Attainment Regions in Feature-Parameter Space for High-Level Debugging in Autonomous Robots](https://mlsanity.com/paper/2108.03150/) (2021)
- [Robust Planning for Autonomous Driving via Mixed Adversarial Diffusion Predictions](https://mlsanity.com/paper/2505.12327/) (2025)

## Abstract (arXiv)

As robots are deployed in complex situations, engineers and end users must develop a holistic understanding of their behaviors, capabilities, and limitations. Some behaviors are directly optimized by the objective function. They often include success rate, completion time or energy consumption. Other behaviors -- e.g., collision avoidance, trajectory smoothness or motion legibility -- are typically emergent but equally important for safe and trustworthy deployment. Designing an objective which optimizes every aspect of robot behavior is hard. In this paper, we advocate for systematic analysis of a wide array of behaviors for holistic understanding of robot controllers and, to this end, propose a framework, RoCUS, which uses Bayesian posterior sampling to find situations where the robot controller exhibits user-specified behaviors, such as highly jerky motions. We use RoCUS to analyze three controller classes (deep learning models, rapidly exploring random trees and dynamical system formulations) on two domains (2D navigation and a 7 degree-of-freedom arm reaching), and uncover insights to further our understanding of these controllers and ultimately improve their designs.