Real-time fault detection and situational awareness for rovers: report on the mars technology program task

Dearden, Richard; Willeke, Thomas; Simmons, Reid; Verma, Vandi; Hutter, Frank; Thrun, Sebastian

doi:10.1109/aero.2004.1367683

Cited by 57 publications

(32 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus f, D, v, a t and az can be updated and calculated by the observations generated in (23) and (4). The estimation errors of the parameters and states, calculated by (23) and (4) are shown from Figure 6 to Figure 10. The corresponding estimated mean errors are specified in Table 1.…”

Section: Simulation Resultsmentioning

confidence: 99%

“…These values are not calculated from (23). The values of parameter particles will be assigned from them.…”

Section: Simulation Resultsmentioning

confidence: 99%

“…In RBPF algorithm, the state space is divided into linear substate space estimated by Kalman filter algorithm and non-linear sub-state space estimated by particle filter algorithm [15][16][17][18]. The RBPF has applications in the hybrid Gaussian [19][20][21], fixed parameters estimation [22], hidden Markov models (HMMs) [20][21], Dirichlet process models and Dynamic Bayesian Networks (DBN) [23]. There is no transition function for the unknown parameter in the state space equations.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Group-SMA Algorithm Based Joint Estimation of Train Parameter and State

Zheng

Han

Kong

et al. 2015

PROMET

View full text Add to dashboard Cite

show abstract

Section: Simulation Resultsmentioning

confidence: 99%

“…These values are not calculated from (23). The values of parameter particles will be assigned from them.…”

Section: Simulation Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Group-SMA Algorithm Based Joint Estimation of Train Parameter and State

Zheng

Han

Kong

et al. 2015

PROMET

View full text Add to dashboard Cite

show abstract

“…Recently, computationally efficient approaches for approximating Bayesian belief using particle filters have been studied as a means for fault detection and identification (Dearden et al 2004;Verma et al 2004;Li and Kadirkamanathan 2001). Particle filters are Monte Carlo methods capable of tracking hybrid state spaces of continuous noisy sensor data and discrete operation states.…”

Section: Related Workmentioning

confidence: 99%

Fault detection in autonomous robots based on fault injection and learning

et al. 2007

View full text Add to dashboard Cite

In this paper, we study a new approach to fault detection for autonomous robots. Our hypothesis is that hardware faults change the flow of sensory data and the actions performed by the control program. By detecting these changes, the presence of faults can be inferred. In order to test our hypothesis, we collect data from three different tasks performed by real robots. During a number of training runs, we record sensory data from the robots while they are operating normally and after a fault has been injected. We use back-propagation neural networks to synthesize fault detection components based on the data collected in the training runs. We evaluate the performance of the trained fault detectors in terms of number of false positives and time it takes to detect a fault. The results show that good fault detectors can be obtained. We extend the set of possible faults and go on to show that a single fault detector can be trained to detect several faults in both a robot's sensors and actuators. We show that fault detectors can be synthesized that are robust to variations in the task, and we show how a fault detector can be trained to allow one robot to detect faults that occur in another robot.

show abstract

“…In these studies, banks of Kalman filters are applied to track multiple models with embedded fault states. Recently, computationally efficient approaches for approximating Bayesian belief using particle filters 1 have been suggested as a means for fault detection and identification [9,10,11].…”

Section: Related Workmentioning

confidence: 99%

Exogenous Fault Detection in a Collective Robotic Task

Christensen

O’Grady

Birattari

et al.

Advances in Artificial Life

View full text Add to dashboard Cite

Abstract. In robotics, exogenous fault detection is the process through which one robot detects faults that occur in other, physically separate robots. In this paper, we study exogenous fault detection in a collective leader-follower task for autonomous robots. We record sensory inputs from the robots while they are operating normally and after simulated faults have been injected. Given that faults are simulated, we can correlate the flow of sensory inputs with the fault state of the robots. We use back-propagation neural networks to synthesize fault detection components. We show that the flow of sensory inputs is sufficient information for performing exogenous fault detection, that is, we show that the leader robot is capable of detecting faults in the follower robot. All results are based on experiments with real robots.

show abstract

Real-time fault detection and situational awareness for rovers: report on the mars technology program task

Cited by 57 publications

References 6 publications

Group-SMA Algorithm Based Joint Estimation of Train Parameter and State

Group-SMA Algorithm Based Joint Estimation of Train Parameter and State

Fault detection in autonomous robots based on fault injection and learning

Exogenous Fault Detection in a Collective Robotic Task

Contact Info

Product

Resources

About