The current study is set up to investigate the problem of planning the trajectories of a multi-robot system. It puts emphasis on the idea of using distributed architecture to plan the trajectories of a group of wheeled mobile robots. Each robot must be able to detect and avoid collision with both static and dynamic obstacles present in its environment/neighborhood. This study seeks to improve the artificial potential field (APF) method in order to have good trajectory planning of a multi-robot system. In order to address the purpose of this work, we employed the hybrid approach, which is a combination of three techniques: the APF method; the neighborhood system; and the notion of priority between the robots. Moreover, the minimum local problem is handled in this paper using the non-minimum speed algorithm. The implemented approach is adapted to solve the trajectory planning problem for a multi-robot system. So, the problem of intersection of robots at the same passage point is solved through using the method of assignment of priority between robots. We also used the neighborhood detection technique to reduce the influence area of each robot and to optimize the time of calculation. The overall system equations associated to the robots are updated at each time of simulation to react at any condition that might suddenly emerge. The approach is implemented with MATLAB/Simulink and Solidworks/Simmechanics.
This work presents a new approach to solve the problem of local minimum which happens usually in the case of multiple robots navigation. The multi-robot path planning based on artificial potential field is among of the most popular method for trajectory planning. However, in some spatial-temporal coordinates, the algebraic sum of all the potentials generated by all the torques/forces is null. In such case, all the robots in the working area stopped and their speeds become null too. Purposely, this work solved the problem of local minimum in a multi-robot system which is validated by Matlab/Simulink simulation.
23 May 2013 in International Journal of Robust and Nonlinear Control
International audienceA new approach to manage actuator redundancy in the presence of faults is proposed based on reliability indicators and a reference governor. The aim is to preserve the health of the actuators and the availability of the system both in the nominal behavior and in the presence of actuator faults. The use of reference governor control allocation is a solution to distribute the control efforts among a redundant set of actuators. In a degraded situation, a reconfigured control allocation strategy is proposed based on on-line re-estimation of the actuator reliability. A benefit of incorporating reliability indicators into over-actuated control system design is the smart management of the redundant actuators and improvement of the system safety. Moreover, when the fault is severe, an adaptation approach using the reference governor is proposed. The reference governor unit is a reference-offset governor based on a discrete-time predictive control strategy. The idea is to modify the reference according to the system constraints, which become stricter after the occurrence of an actuator fault. The proposed approach is illustrated with a flight control application
Wireless Sensor Networks are deployed in harsh environments. Their key advantage is there flexibility and low cost. But they can face many failures which created the need to improve data accuracy. Many artificial intelligence techniques has demonstrated impressive results in fault detection and faults diagnosis. Lately, machine learning emerged as a powerfull artificial intelligence based technique to solve the problem of failures in WSN. In this paper, a multi-fault classification is evaluated using deep learning technique based on LSTM classifier and then compared with different machine learning techniques such as Support Vector Machine (SVM), Random Forest (RF), Multilayer Perceptron (MLP)and Probabilistic Neural Network (PNN). The performance of this mentioned techniques used for fault detection in WSNs were compared based on four metrics: Detection Accuracy (DA), True Positive Rate (TPR), Matthews Correlation Coefficients (MCC)and False Alarm (FA).