Recovering from External Disturbances in Online Manipulation through State-Dependent Revertive Recovery Policies

Abstract

Robots are increasingly entering uncertain and unstructured environments. Within these, robots are bound to face unexpected external disturbances like accidental human or tool collisions. Robots must develop the capacity to respond to unexpected events. That is not only identifying the sudden anomaly, but also deciding how to handle it. In this work, we contribute a recovery policy that allows a robot to recovery from various anomalous scenarios across different tasks and conditions in a consistent and robust fashion. The system organizes tasks as a sequence of nodes composed of internal modules such as motion generation and introspection. When an introspection module flags an anomaly, the recovery strategy is triggered and reverts the task execution by selecting a target node as a function of a state dependency chart. The new skill allows the robot to overcome the effects of the external disturbance and conclude the task. Our system recovers from accidental human and tool collisions in a number of tasks. Of particular importance is the fact that we test the robustness of the recovery system by triggering anomalies at each node in the task graph showing robust recovery everywhere in the task. We also trigger multiple and repeated anomalies at each of the nodes of the task showing that the recovery system can consistently recover anywhere in the presence of strong and pervasive anomalous conditions. Robust recovery systems will be key enablers for long-term autonomy in robot systems. Supplemental information including videos, code, and result analysis can be found at [1].

Fast, robust, and versatile event detection through HMM belief state gradient measures

“…This is a critical point in assessing the robustness, accuracy, versatility, and reaction speed of the technique, as conditions can change drastically in a post recovery environment from that used in training for the original identification tasks. In [9], skill identification and anomaly detection were implemented through nonparametric HMM models. A generic recovery system was implemented and event detection studied after recovery actions.…”

“…recovery behavior, we count how many false-positives are triggered before moving to the next skill execution. The robot recovery behavior is detailed in [9]. Five nominal and five anomalous trials are used for the analysis.…”

“…With increasing computation, sensor, and actuator resources, continuous multi-modal signals are used along with better robot and environmental modeling for better event detection [5]- [10]. Recently, there has been interest in not just identifying anomalies but also incrementally classifying them and integrating recovery mechanisms into their systems [9], [11]. The goal is to close loops between high-level event detectors and low-level robot controllers that can optimally adjust or recover from anomalous events.…”

“…The goal is to close loops between high-level event detectors and low-level robot controllers that can optimally adjust or recover from anomalous events. Not much work has been done in the study of event detection in the post-anomaly-recovery stage of a task [9], [12]. Little quantitative and qualitative analysis is available for such techniques, which are critical for long-term autonomy.…”

A Latent State-Based Multimodal Execution Monitor with Anomaly Detection and Classification for Robot Introspection

“…For the anomaly detection, we achieved it by comparing the cumulative log-likelihood deviation to a given threshold from nominal executions [1,2,9]. An improved anomaly detector will be proposed in this paper, which can more effectively and robustly detect the anomalies.…”