Neural networks have become state-of-the-art for computer vision problems because of their ability to efficiently model complex functions from large amounts of data. While neural networks can be shown to perform well empirically for a variety of tasks, their performance is difficult to guarantee. Neural network verification tools have been developed that can certify robustness with respect to a given input image; however, for neural network systems used in closed-loop controllers, robustness with respect to individual images does not address multi-step properties of the neural network controller and its environment. Furthermore, neural network systems interacting in the physical world and using natural images are operating in a black-box environment, making formal verification intractable. This work combines the adaptive stress testing (AST) framework with neural network verification tools to search for the most likely sequence of image disturbances that cause the neural network controlled system to reach a failure. An autonomous aircraft taxi application is presented, and results show that the AST method finds failures with more likely image disturbances than baseline methods. Further analysis of AST results revealed an explainable cause of the failure, giving insight into the problematic scenarios that should be addressed.
University of Michigan（密歇根大学）；Plus.ai Inc.（智加科技公司）；California Institute of Technology（加州理工学院）；Visteon Corporation（伟世通公司）
Accomplishing safe and efficient driving is one of the predominant challenges in the controller design of connected automated vehicles (CAVs). It is often more convenient to address these goals separately and integrate the resulting controllers. In this study, we propose a controller integration scheme to fuse performance-based controllers and safety-oriented controllers safely for the longitudinal motion of a CAV. The resulting structure is compatible with a large class of controllers, and offers flexibility to design each controller individually without affecting the performance of the other…
In the control of bipedal locomotion, linear velocity of the center of mass has been widely accepted as a primary variable for summarizing a robot's state vector. The ubiquitous massless-legged linear inverted pendulum (LIP) model is based on it. In this paper, we argue that angular momentum about the contact point has several properties that make it superior to linear velocity for feedback control. So as not to confuse the benefits of angular momentum with any other control design decisions, we first reformulate the standard LIP controller in terms of angular momentum. We then implement the r…
This paper develops a Closed-Loop Error Learning Control (CLELC) algorithm for feedback linearizable systems with experimental validation on a mobile robot. Traditional feedback and feedforward controllers are designed based on the nominal model by using Feedback Linearization Control (FLC) method. Then, an intelligent controller is designed based on sliding mode learning algorithm that utilizes closed-loop error dynamics to learn the system behavior. The controllers are working in parallel, and the intelligent controller can gradually replace the feedback controller from the control of the sy…
Validation is a key challenge in the search for safe autonomy. Simulations are often either too simple to provide robust validation, or too complex to tractably compute. Therefore, approximate validation methods are needed to tractably find failures without unsafe simplifications. This paper presents the theory behind one such black-box approach: adaptive stress testing (AST). We also provide three examples of validation problems formulated to work with AST.
Roundabouts in conjunction with other traffic scenarios, e.g., intersections, merging roadways, speed reduction zones, can induce congestion in a transportation network due to driver responses to various disturbances. Research efforts have shown that smoothing traffic flow and eliminating stop-and-go driving can both improve fuel efficiency of the vehicles and the throughput of a roundabout. In this paper, we validate an optimal control framework developed earlier in a multi-lane roundabout scenario using the University of Delaware's scaled smart city (UDSSC). We first provide conditions where…
Stanford University（斯坦福大学）；University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Determining possible failure scenarios is a critical step in the evaluation of autonomous vehicle systems. Real-world vehicle testing is commonly employed for autonomous vehicle validation, but the costs and time requirements are high. Consequently, simulation-driven methods such as Adaptive Stress Testing (AST) have been proposed to aid in validation. AST formulates the problem of finding the most likely failure scenarios as a Markov decision process, which can be solved using reinforcement learning. In practice, AST tends to find scenarios where failure is unavoidable and tends to repeatedly…
During the development of autonomous systems such as driverless cars, it is important to characterize the scenarios that are most likely to result in failure. Adaptive Stress Testing (AST) provides a way to search for the most-likely failure scenario as a Markov decision process (MDP). Our previous work used a deep reinforcement learning (DRL) solver to identify likely failure scenarios. However, the solver's use of a feed-forward neural network with a discretized space of possible initial conditions poses two major problems. First, the system is not treated as a black box, in that it requires…
Lyapunov-Based PI-Like Control for Robust Trajectory Tracking of a Four-Wheel Independently Driven and Steered Robot: Design and Experimental Validation
Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb（Zagreb大学机械工程与造船工程学院）；Regional Centre of Excellence for Robotic Technology (CRTA)（机器人技术卓越研究中心）；Croatian Academy of Sciences and Arts（克罗地亚科学院）
CReSTIC, IUT de Troyes（特鲁瓦大学技术学院CReSTIC实验室）；Mines ParisTech, PSL Research University（巴黎高等矿业学院，PSL研究大学）；École polytechnique（巴黎综合理工学院）；AL.I.E.N.（AL.I.E.N.公司）；IFSTTAR（法国法国交通发展与网络科学研究院）；CNRS（法国国家科学研究中心）；Supélec（高等电力学院）；Université Paris-Sud（巴黎第十一大学）
A new model-free setting and the corresponding "intelligent" P and PD controllers are employed for the longitudinal and lateral motions of a vehicle. This new approach has been developed and used in order to ensure simultaneously a best profile tracking for the longitudinal and lateral behaviors. The longitudinal speed and the derivative of the lateral deviation, on one hand, the driving/braking torque and the steering angle, on the other hand, are respectively the output and the input variables. Let us emphasize that a "good" mathematical modeling, which is quite difficult, if not impossible …
AdapJ: An Adaptive Extended Jacobian Controller for Soft Manipulators
Biorobotics Institute and the Department of Excellence in Robotics and AI, Scuola Superiore Sant’Anna（生物机器人研究所和机器人与人工智能卓越部门，圣安娜高等学院）；Multi-scale Medical Robotics Centre and Chow Yuk Ho Technology Centre for Innovative Medicine, The Chinese University of Hong Kong（多尺度医疗机器人中心和创新医学技术中心，香港中文大学）；Department of computer systems, Tallinn University of Technology（计算机系统系，塔林技术大学）
Robots are ideal for open-pit mining on the Moon as its a dull, dirty, and dangerous task. The challenge is to scale up productivity with an ever-increasing number of robots. This paper presents a novel method for developing scalable controllers for use in multi-robot excavation and site-preparation scenarios. The controller starts with a blank slate and does not require human-authored operations scripts nor detailed modeling of the kinematics and dynamics of the excavator. The 'Artificial Neural Tissue' (ANT) architecture is used as a control system for autonomous robot teams to perform resou…
LAAS-CNRS（法国国家科学研究中心系统分析与架构实验室）；Université de Toulouse（图卢兹大学）；CNRS（法国国家科学研究中心）
GenoM is an approach to develop robotic software components, which can be controlled, and assembled to build complex applications. Its latest version GenoM3, provides a template mechanism which is versatile enough to deploy components for different middleware without any change in the specification and user code. But this same template mechanism also enables us to automatically synthesize formal models (for two Validation and Verification frameworks) of the final components. We illustrate our approach on a real deployed example of a drone flight controller for which we prove offline real-time …
