This paper presents an automated knee rehabilitation system designed to assist patients during the gait cycle. The system utilizes a predictive model based on a neural network to generate trajectories of the knee motion from data captured by a motion capture system. Also, some control schemes, a PID, a Second Order Sliding Mode Control, and an Active Disturbance Rejection Control, were implemented to follow these trajectories, considering knee dynamics and external disturbances. Finally, the experimental setup was developed to implement and compare these controllers. This adaptable solution has potential applications for post-surgical patients, athletes, and children with motor difficulties.

This paper proposes a fluid based macroscopical perspective for modeling and simulation of Vehicular Traffic Networks based on the connection of continuous dynamical agents. The agents' flows are designed to be facilitated by a high-level planner that represents conventional traffic signals cycles. The research examines the dynamics of groups using various state-space representations. Some numerical simulations are presented to demonstrate the approach's use and scalability. In order to evaluate the evolution of the system design, several performance criteria were also established and tested.

This paper presents a novel modeling of complex urban traffic networks based on the interconnections of agents which represent the vehicular fluid dynamics in physical spaces like streets and intersections, supervised by a Petri Net. The agent's behavior is ruled by simple nonlinear differential equations based on the vehicular density, average vehicle velocities, congestions and occupation capacities. The Petri Net is designed to manage the traffic light cycles, enabling or disabling the flow of vehicles to and from adjacent agents. The main contribution of the paper is to define the basis of a hybrid and modular system to be addressed by the theory of multi-agent systems and supervisory control of Petri Nets. An example modeling a real intersection in the city of Montevideo is presented. The model's behavior is compared with a common traffic simulator in order to show the potential of the approach and its scalability to more complex traffic networks.

An analysis of the convergence and non-collision conditions of a formation control strategy for multi-agent robots based on local potential functions is presented. The goal is to coordinate a group of agents, considered as points in plane, to achieve a particular formation. The control law is designed using local attractive forces only where every agent knows the position of another two agents reducing the requirements of the control law implementation. The control law guarantees the convergence to the desired formation but does not avoid inter-agent collisions. A set of necessary and sufficient non-collision conditions based on the explicit solution of the closed-loop system is derived. The conditions allow to conclude from the initial conditions whether or not the agents will collide. The formal proof is presented for the case of three agents. The result is extended to the case of formations of three unicycles.