AI-based Control Approaches for Multiple Mobile Robots  Project

 The project SEP-CONACYT-ANUIES-ECOS NORD 315597 (2021-2024):


 “Artificial Intelligence–based Control Approaches for Multiple Mobile Robots” 


is a bilateral research collaboration between the Tecnológico Nacional de México/I.T. La Laguna, Mx., and the Inria Lille-Nord Europe, Fr. The project is focused on studying several tracking tasks for autonomous mobile robot systems, particularly unmanned aerial vehicles (UAV) and wheeled mobile robots (WMR). This project aims to develop robust control and navigation schemes by combining the methods of artificial intelligence and control theory. The objective is to solve different collective tasks in autonomous mobile robot systems. The project also comprises students and researchers from the CNAM (Conservatoire national des arts et métiers), ESIME-IPN, CITEDI-IPN, and ITESM Guadalajara. 

This project tries to address an international and national scope problem in engineering. 

Participants



Publications


JCR Journals


Book Chapters


International Conferences


National Conferences


PhD Thesis



Courses



Research Seminars

 


Research Stays


Researchers


Students

Overview of the Problem

Analysis and implementation of control, and localization algorithms are required for different Autonomous Mobile Robots Systems (AMoRS) in a great variety of applications due to their low costs, social impact, and human life protection. The study of AMoRS, as well as its applications, has increased exponentially in recent years. For instance, the UAVs are used for supervision and maintenance in electric power distribution networks, wind turbines or generators, hydroelectric plants, and solar cells; all of them located at considerable heights and at difficult access zones for humans; while the WMRs are used in the industrial space, in medical/surgical applications, in painting and de-painting applications, accessing areas dangerous to humans, etc. Both of them are also used to study and reproduce the behaviour of different multi-agent systems in nature, and they can also work together; for instance, in explorer missions, where there is a need to charge their batteries from time to time. The WMRs play the role of the docking stations while the UAVs realize the exploration. 

In Mexico and France, many of the applications are focused on supervision and maintenance of electric power networks. Nevertheless, in the research field, there are still many problems to be solved to add safety, robustness and efficiency, to the UAVs that are employed in those applications.

All those tasks require to study the dynamics of the different AMoRS, to formally analyse the different application problems, to design different control and path-planning schemes, and to implement different vision algorithms that are able to ensure the task realization safely and efficiently. Moreover, all these analysis and design problems for AMoRS become more complex if one takes into account uncertainty: lack of measurement accuracy, obstacles, and external disturbances that appear during the different tasks. In addition, it is wroth to highlight the importance of cooperation for mobile robots. Application of the multiple AMoRS offers a solution to the following issues:

Additionally, the AMoRS provide some other advantages:

In spite of all the mentioned applications, there are still many open problems related to the design of robust navigation schemes and vision algorithms implementation that are able to realize the corresponding tasks in a safe and efficient way; under the presence of uncertainties in the mathematical models, lack of measurement accuracy and external disturbances such as wind gusts, irregular surfaces or obstacles. In addition to this, it is very important that the tasks are realized in an autonomous way and that the AMoRS are capable of dealing with the above-mentioned problems without the need for re-configurations and human intervention, i.e., requirements of vision algorithms based on AI.

Moreover, it is important to highlight that most of the mentioned AMoRS applications may be essentially seen, in the control theory framework, as robust trajectory planning, consensus, formation and obstacle avoidance problems for single or multiple vehicles, i.e., multi-agents or multi AMoRS. Nevertheless, the solution to these problems is not straightforward and a lot of research effort must be made in both areas, control theory and AI-based algorithms.

Homogeneous Controller – Trajectory Tracking in Unicycle Mobile Robots – Experimental Results QBot2

A homogeneous controller is developed based on a particular cascade control strategy. The design is based on the canonical homogeneous norm and the degree of homogeneity. Some experimental results illustrate the performance of the proposed homogeneous control in the UMR QBot2 by Quanser.

The trajectory–tracking experiments, which consider soil on the surface, illustrate the performance of the proposed homogeneous controller (HC) compared with two other controllers, i.e., a first–order sliding–mode (FOSM) controller proposed in [1] and a nonlinear controller (NC) presented in [2]. 

[1] M. Mera, H. Ríos, and E. A Martínez. A sliding–mode–based controller for trajectory tracking of perturbed unicycle mobile robots. Control Engineering Practice, 102:104548, 2020.

[2] M. Maghenem, A. Loría, and E. Panteley. Formation-tracking control of autonomous vehicles under relaxed persistency of excitation conditions. IEEE Transactions on Control Systems Technology, 26(5):1860–1865, 2017. 

A Super-Twisting-based Controller for Trajectory Tracking of Perturbed Unicycle Mobile Robots

In this video, a robust position tracking control problem for a unicycle mobile robot is presented. To this aim, a Super-Twisting-based Controller is designed. The results show the performance for different desired trajectories. For more information, check out the cite at the end of this description. 

En este video se presentan el problema de seguimiento de trayectorias para un robot móvil tipo uniciclo. Para este propósito se diseña un controlador basado en el algoritmo Super-Twisting. Los resultados muestran el desempeño para diferentes trayectorias deseadas. 

P. Rochel, H. Ríos, M. Mera and A. Dzul. “Trajectory Tracking for Uncertain Unicycle Mobile Robots: A Super–Twisting Approach”. Control Engineering Practice 122, 2022, 105078.  

Fault Accommodation Control for Trajectory Tracking in Quad-Rotors

An actuator fault accommodation controller is developed to solve the trajectory tracking problem in Quad-Rotors under the effects of faults in multiple actuators and external disturbances. 

The proposed fault accommodation approach is composed of a fault identification module and a baseline robust nominal controller. The fault identification module is based on a finite-time sliding-mode observer that provides a set of residuals using only the output information. The fault accommodation strategy uses fault identification to partially compensate the actuator faults allowing the usage of a baseline robust-nominal controller that deals with external disturbances.

Romeo Falcón, Héctor Ríos, Alejandro Dzul “An Actuator Fault Accommodation Sliding-Mode Control Approach for Trajectory Tracking in Quad-Rotors”. In the 60th IEEE Conference on Decision and Control (CDC), Austin, Texas, USA, 2021, pp. 7100–7105.