Global methodology in control architecture to improve mobile robot reliability

Abstract

This paper presents a global methodology developed to increase the reliability of mobile robots. An initial analysis of robot functions and their corresponding significant failures enabled us to introduce dedicated observation modules in the embedded architecture to monitor fault events. Once detected, the functioning mode of the robot can be adapted according to the failure severity level to ensure the success of its mission. The methodology was applied in a case study and the experimental implementation and results are detailed.

Assessing Critical Adaptations in Automated Adaptive Software Systems by Stage Decomposition

“…Although previous studies referred to decision-making transfers in adjustable autonomy [23][24][25][26][27], they do not explicitly assess the impact of safety-critical adaptations on the assets. Some studies considered the risk of adjustable autonomy [18], [19], [26], [28][29][30][31]. Bush et al defined risk as the degree that the mission goal cannot be achieved [28].…”

“…Bush et al defined risk as the degree that the mission goal cannot be achieved [28]. Durand et al proposed a formal modeling method to detect inconsistency of control architecture, however, the method does not consider the critical impact of the inconsistency on the assets [29]. Roehr and Shi defined the highest risk in a reconfigurable integrated multi-robot exploration adaptive system as the loss of the robot.…”

Fault tolerance enhancement using autonomy adaptation for autonomous mobile robots

“…In spite of numerous works concerning dependability concepts for autonomous robots, there is a lack of global and structured approach including all these aspects of fault tolerance. We have proposed in [19] a methodology aiming to address, in a flexible and generic way, all these aspects. It is based on four successive steps allowing the enhancement of robots reliability: fault identification, fault detection and diagnosis, and fault recovery.…”

“…To enhance reliability and robustness of robotic systems, we propose to integrate in this architecture fault tolerant mechanisms for fault detection and fault recovery. These mechanisms are included in a global methodology detailed in [19].…”

Mobile Robot Lab Project to Introduce Engineering Students to Fault Diagnosis in Mechatronic Systems

“…Mobile robots are another significant example of safetycritical autonomous system that should be designed to tolerate failing execution and uncertainty [16] [17]. Precisely, the reliability demanded by challenging applications [18] has fostered a growing research interest in FDD at different operation levels: steering mechanisms [19], anomalous actuator and wheel response [20], collision detection in spite of sensor limitations [21], localization failures due to abrupt wheel slippage [22], multi-robot performance [23], and monitoring of task states and mission goals [24].…”