Neural-Dynamic Optimization-Based Model Predictive Control for Tracking and Formation of Nonholonomic Multirobot Systems

Abstract

In this paper, a neural-dynamic optimization-based nonlinear model predictive control (NMPC) is developed for the multiple nonholonomic mobile robots formation. First, a model-based monocular vision method is developed to obtain the location information of the leader. Then, a separation-bearing-orientation scheme (SBOS) control strategy is proposed. During the formation motion, the leader robot is controlled to track the desired trajectory and the desired leader-follower relationship can be maintained through the SBOS method. Finally, the model predictive control (MPC) is utilized to maintain the desired leader-follower relationship. To solve the MPC generated constrained quadratic programming problem, the neural-dynamic optimization approach is used to search for the global optimal solution. Compared to other existing formation control approaches, the proposed solution is that the NMPC scheme exploit prime-dual neural network for online optimization. Finally, by using several actual mobile robots, the effectiveness of the proposed approach has been verified through the experimental studies.

Precise Measurement of Position and Attitude Based on Convolutional Neural Network and Visual Correspondence Relationship

“…A measurement scheme with a defined cooperation goal can significantly enhance measurement universality [27]. For this reason, we use the form of cooperation objectives in the program for adequate training and testing.…”

Model Predictive Control of Robotic Grinding Based on Deep Belief Network

“…Robotic Grinding Trajectory Prediction. Substitute (19) and 15into (14) to obtain the relationship between robotic grinding force deviation and feed rate…”

“…Some scholars try to improve nonlinear model predictive control approach by using intelligent algorithms such as neural network [11][12][13]. Li [14] proposed a nonlinear model control method based on neural dynamic network, where the neural dynamic network is used to obtain optimal values of the formulated 2 Complexity constrained quadratic programming (QP) problem derived from the cost function of nonlinear model predictive control model. Zeng [15] used Gaussian radial basis function (RBF) neural networks to improve the nonlinear model predictive control approach applied in the control of nonlinear multivariable systems.…”

“…Rewrite(15) according to(14) ( +1 ) − * ( +1 ) = ( ) − * ( ) + ( ( )) ( +1 ) − ( ) = * ( +1 ) − * ( ) + ( ( )) Δ ( ) = Δ * ( ) + ( ( )) ) = ( +1 ) − ( ),̇( ) = ( ), Δ ( ) = ( ) . According to (6) there is Δ ( ) = ( +1 )− ( ) = ( +1 ) − ( +1 ), and (17) can be rewritten as Δ ( ) = Δ̈( ) + ( + ) Δ ( )…”

Formation Control for a Fleet of Autonomous Ground Vehicles: A Survey

“…In this study, a follower robot was tracking the features of a leader robot through a camera and two controllers were developed: a formation controller to maintain formation and a camera controller to provide visual measurements. Furthermore, a vision-based leader-follower formation control was achieved by developing a neural-dynamic optimization-based nonlinear model predictive control (MPC) [27]. In this study, a camera on follower robot was employed to track the features and to measure state and velocity of the leader robot.…”