Reinforcement Learning for Safety Testing: Lessons from A Mobile Robot Case Study

Safety-critical robot systems need thorough testing to expose design flaws and software bugs which could endanger humans. Testing in simulation is becoming increasingly popular, as it can be applied early in the development process and does not endanger any real-world operators. However, not all safety-critical flaws become immediately observable in simulation. Some may only become observable under certain critical conditions. If these conditions are not covered, safety flaws may remain undetected. Creating critical tests is therefore crucial. In recent years, there has been a trend towards using Reinforcement Learning (RL) for this purpose. Guided by domain-specific reward functions, RL algorithms are used to learn critical test strategies. This paper presents a case study in which the collision avoidance behavior of a mobile robot is subjected to RL-based testing. The study confirms prior research which shows that RL can be an effective testing tool. However, the study also highlights certain challenges associated with RL-based testing, namely (i) a possible lack of diversity in test conditions and (ii) the phenomenon of reward hacking where the RL agent behaves in undesired ways due to a misalignment of reward and test specification. The challenges are illustrated with data and examples from the experiments, and possible mitigation strategies are discussed.

Uncertainty-aware Risk Assessment of Robotic Systems via Importance Sampling

In this paper, we introduce a probabilistic approach to risk assessment of robot systems by focusing on the impact of uncertainties. While various approaches to identifying systematic hazards (e.g., bugs, design flaws, etc.) can be found in current literature, little attention has been devoted to evaluating risks in robot systems in a probabilistic manner. Existing methods rely on discrete notions for dangerous events and assume that the consequences of these can be described by simple logical operations. In this work, we consider measurement uncertainties as one main contributor to the evolvement of risks. Specifically, we study the impact of temporal and spatial uncertainties on the occurrence probability of dangerous failures, thereby deriving an approach for an uncertainty-aware risk assessment. Secondly, we introduce a method to improve the statistical significance of our results: While the rare occurrence of hazardous events makes it challenging to draw conclusions with reliable accuracy, we show that importance sampling -- a technique that successively generates samples in regions with sparse probability densities -- allows for overcoming this issue. We demonstrate the validity of our novel uncertainty-aware risk assessment method in three simulation scenarios from the domain of human-robot collaboration. Finally, we show how the results can be used to evaluate arbitrary safety limits of robot systems.

Safety Evaluation of Robot Systems via Uncertainty Quantification

In this paper, we present an approach for quantifying the propagated uncertainty of robot systems in an online and data-driven manner. Especially in Human-Robot Collaboration, keeping track of the safety compliance during run time is essential: Misclassifying dangerous situations as safe might result in severe accidents. According to official regulations (eg, ISO standards), safety in industrial robot applications depends on critical parameters, such as the distance and relative velocity between humans and robots. However, safety can only be assured given a measure for the reliability of these parameters. While different risk detection and mitigation approaches exist in literature, a measure that can be used to evaluate safety limits online, and succinctly implies whether a situation is safe or dangerous, is missing to date. Motivated by this, we introduce a generalizable method for calculating the propagated measurement uncertainty of arbitrary parameters, that captures the accumulated uncertainty originating from sensory devices and environmental disturbances of the system. To show that our approach delivers correct results, we perform validation experiments in simulation. In addition, we employ our method in two real-world settings and demonstrate how quantifying the propagated uncertainty of critical parameters facilitates assessing safety online in Human-Robot Collaboration.

Hazard Analysis of Collaborative Automation Systems: A Two-layer Approach based on Supervisory Control and Simulation

Safety critical systems are typically subjected to hazard analysis before commissioning to identify and analyse potentially hazardous system states that may arise during operation. Currently, hazard analysis is mainly based on human reasoning, past experiences, and simple tools such as checklists and spreadsheets. Increasing system complexity makes such approaches decreasingly suitable. Furthermore, testing-based hazard analysis is often not suitable due to high costs or dangers of physical faults. A remedy for this are model-based hazard analysis methods, which either rely on formal models or on simulation models, each with their own benefits and drawbacks. This paper proposes a two-layer approach that combines the benefits of exhaustive analysis using formal methods with detailed analysis using simulation. Unsafe behaviours that lead to unsafe states are first synthesised from a formal model of the system using Supervisory Control Theory. The result is then input to the simulation where detailed analyses using domain-specific risk metrics are performed. Though the presented approach is generally applicable, this paper demonstrates the benefits of the approach on an industrial human-robot collaboration system.

SpeedFolding: Learning Efficient Bimanual Folding of Garments

Folding garments reliably and efficiently is a long standing challenge in robotic manipulation due to the complex dynamics and high dimensional configuration space of garments. An intuitive approach is to initially manipulate the garment to a canonical smooth configuration before folding. In this work, we develop SpeedFolding, a reliable and efficient bimanual system, which given user-defined instructions as folding lines, manipulates an initially crumpled garment to (1) a smoothed and (2) a folded configuration. Our primary contribution is a novel neural network architecture that is able to predict pairs of gripper poses to parameterize a diverse set of bimanual action primitives. After learning from 4300 human-annotated and self-supervised actions, the robot is able to fold garments from a random initial configuration in under 120s on average with a success rate of 93%. Real-world experiments show that the system is able to generalize to unseen garments of different color, shape, and stiffness. While prior work achieved 3-6 Folds Per Hour (FPH), SpeedFolding achieves 30-40 FPH.

Learning Time-optimized Path Tracking with or without Sensory Feedback

In this paper, we present a learning-based approach that allows a robot to quickly follow a reference path defined in joint space without exceeding limits on the position, velocity, acceleration and jerk of each robot joint. Contrary to offline methods for time-optimal path parameterization, the reference path can be changed during motion execution. In addition, our approach can utilize sensory feedback, for instance, to follow a reference path with a bipedal robot without losing balance. With our method, the robot is controlled by a neural network that is trained via reinforcement learning using data generated by a physics simulator. From a mathematical perspective, the problem of tracking a reference path in a time-optimized manner is formalized as a Markov decision process. Each state includes a fixed number of waypoints specifying the next part of the reference path. The action space is designed in such a way that all resulting motions comply with the specified kinematic joint limits. The reward function finally reflects the trade-off between the execution time, the deviation from the desired reference path and optional additional objectives like balancing. We evaluate our approach with and without additional objectives and show that time-optimized path tracking can be successfully learned for both industrial and humanoid robots. In addition, we demonstrate that networks trained in simulation can be successfully transferred to a real Kuka robot.

Testing Robot System Safety by creating Hazardous Human Worker Behavior in Simulation

We introduce a novel simulation-based approach to identify hazards that result from unexpected worker behavior in human-robot collaboration. Simulation-based safety testing must take into account the fact that human behavior is variable and that human error can occur. When only the expected worker behavior is simulated, critical hazards can remain undiscovered. On the other hand, simulating all possible worker behaviors is computationally infeasible. This raises the problem of how to find interesting (i.e., potentially hazardous) worker behaviors given a limited number of simulation runs. We frame this as a search problem in the space of possible worker behaviors. Because this search space can get quite complex, we introduce the following measures: (1) Search space restriction based on workflow-constraints, (2) prioritization of behaviors based on how far they deviate from the nominal behavior, and (3) the use of a risk metric to guide the search towards high-risk behaviors which are more likely to expose hazards. We demonstrate the approach in a collaborative workflow scenario that involves a human worker, a robot arm, and a mobile robot.

Learning a Generative Transition Model for Uncertainty-Aware Robotic Manipulation

Robot learning of real-world manipulation tasks remains challenging and time consuming, even though actions are often simplified by single-step manipulation primitives. In order to compensate the removed time dependency, we additionally learn an image-to-image transition model that is able to predict a next state including its uncertainty. We apply this approach to bin picking, the task of emptying a bin using grasping as well as pre-grasping manipulation as fast as possible. The transition model is trained with up to 42000 pairs of real-world images before and after a manipulation action. Our approach enables two important skills: First, for applications with flange-mounted cameras, picks per hours (PPH) can be increased by around 15% by skipping image measurements. Second, we use the model to plan action sequences ahead of time and optimize time-dependent rewards, e.g. to minimize the number of actions required to empty the bin. We evaluate both improvements with real-robot experiments and achieve over 700 PPH in the YCB Box and Blocks Test.

Jerk-limited Real-time Trajectory Generation with Arbitrary Target States

We present Ruckig, an algorithm for Online Trajectory Generation (OTG) respecting third-order constraints and complete kinematic target states. Given any initial state of a system with multiple Degrees of Freedom (DoFs), Ruckig calculates a time-optimal trajectory to an arbitrary target state defined by its position, velocity, and acceleration limited by velocity, acceleration, and jerk constraints. The proposed algorithm and implementation allows three contributions: (1) To the best of our knowledge, we derive the first OTG algorithm with non-zero target acceleration, resulting in a complete defined target state. (2) This is the first open-source prototype of time-optimal OTG with limited jerk. (3) Ruckig allows for directional velocity and acceleration limits, enabling robots to better use their dynamical resources. We evaluate the robustness and real-time capability of the proposed algorithm on a test suite with over 1,000,000,000 random trajectories as well as in real-world applications.

Robot Learning of 6 DoF Grasping using Model-based Adaptive Primitives

Robot learning is often simplified to planar manipulation due to its data consumption. Then, a common approach is to use a fully-convolutional neural network to estimate the reward of grasp primitives. In this work, we extend this approach by parametrizing the two remaining, lateral Degrees of Freedom (DoFs) of the primitives. We apply this principle to the task of 6 DoF bin picking: We introduce a model-based controller to calculate angles that avoid collisions, maximize the grasp quality while keeping the uncertainty small. As the controller is integrated into the training, our hybrid approach is able to learn about and exploit the model-based controller. After real-world training of 27000 grasp attempts, the robot is able to grasp known objects with a success rate of over 92% in dense clutter. Grasp inference takes less than 50ms. In further real-world experiments, we evaluate grasp rates in a range of scenarios including its ability to generalize to unknown objects. We show that the system is able to avoid collisions, enabling grasps that would not be possible without primitive adaption.