Contrôle

Now showing 1 - 5 of 5

Feedback motion planning for car-like robots
(Université M'hamed Bougara Boumerdès: Institue de génie electronic et electric, 2024) Derar, Iyed; Khoumari, Malik; Guernane, Reda (supervisor)
In this study, we address the problem of feedback motion planning for a car-like robot de-scribed by a kinematic model, navigating through obstacles. For this purpose, we designed two sampling-based algorithms equipped with the feedback property, a Modifie dRapidly Exploring Random Tree Star (RRT*) algorithm and a Funnel-Graph algorithm. Both al-gorithms generate collision-free paths while accounting for non-holonomic constraints and uncertainties. These generated paths are then fed to the pure pursuit controller which han-dles the robot motion execution. Our approach is validated by testing the algorithms’ ability to handle differen tuncertaintie san dadaptability to environmental changes,including a performance comparison between them. The results show the superiority of the funnel-graph algorithm across all tests, which makes it best suited for real-time applications.
Feedback motion planning with simulation based LQR-trees
(2021) Chadli, Kouider; Guernane, Reda (supervisor)
In autonomous and non-autonomous systems, a motion planner generates reference trajectories which are tracked by a low-level controller. In this report we consider the problem of generating a feedback motion planning algorithm for a nonlinear dynamical systems; the algorithm computes the stability regions to build a set of LQR-stabilized trajectories by generating a feedback control law from a set of initial conditions that are goal reachable. Furthermore, we consider the case where these plans must be generated offline, because the LQR trees lack the ability to handle events in which the goal and environments are unknown till run-time. Moreover, the algorithm approximates the funnel [2] of a trajectory using the one step Lyapunov method which is a sampling-based approach, generating a control law that stabilizes the bounded set to the goal is equivalent to adding trajectories to the tree until their funnels cover the design set. We further validate our approach by carefully evaluating the guarantees on invariance provided by funnels on nonlinear systems. We demonstrate and validate our method using simulation experiments of some nonlinear models. These demonstrations constitute examples of provably safe and robust control for robotic systems with complex nonlinear dynamics with Obstacles and dynamic constraints.
Feedback motion planning for teyhered mobile robots
(2021) Hambli, Billel; Charef, Redha; Guernane, Reda (supervisor)
In this report titled Feedback motion planning for tethered mobile robots, we employ many concepts in order to arrive at a methodology by which path planning of tethered robots can be achieved using feedback. These concepts include constructing a map of the encountered homotopic classes in our environment, building an augmented and tether aware virtual potential field that is responsible for both the advancement towards the goal and the retraction to the anchor of the tether, and the use of path shortening and length calculation algorithms. For the purpose of simplicity, our approach is only limited to the cases where there is no tether crossing while being wrapped around an obstacle. A discrete implementation of the suggested strategy using wavefront planner is presented as a proof of concept.
Kinodynamic planning for omnidirectional mobile robot
(2020) Halitmi, Hiba; Guernane, Reda (supervisor)
This project is about kinodynamic planning for a three-wheeled omnidirectional mobile robot in a static environment. First we present the kinematics and dynamics of the robot, then we use those models to plan the trajectory of the robot in the state space using the RRT then the anytime RRT sampling-based algorithms. The algorithms are then implemented and tested on MATLAB.
End-to-End learning-based navigation of autonoumous mobile robot
(2020) Mehrab, Anis Abdeldjalil; Guernane, Reda (supervisor)
In this work we present an end-to-end learning approach that is able to perform target- oriented navigation and collision avoidance using Deep Neural Network. This approach can be defined as learning a model that maps sensory inputs, such as raw 2D-laser range findings and a target position, to navigation actions for controlling the mobile robot such as steering commands. Compared to the traditional autonomous navigation systems, which often require perception, localization, mapping, and path planning, the end-to-end learning approach offers a more efficient method. which utilize large set of expert navigation demonstrations to learn the desired navigation policy. The end-to-end learning approach has gained considerable interests in autonomous navigation in academic and industrial fields. Researches have already used different artificial neural networks to predict steering commands. However, most of the existing end-to-end methods are used for lane keeping for self-driving cars. therefore, we propose an end-to-end navigation model for mobile robots that is based on a Convolutional Neural Network (CNN). The network was trained using expert demonstration data which was generated in virtual simulation environments. The learned model was test in real time simulation and gave an acceptable result, however, it suffered when it encounters situations that requires hard maneuvers. Therefore, in order to overcome some of these difficulties, we proposed an improved model which incorporates the temporal information in the prediction process using the Long Short-Term Memory (LSTM) network. basically, this model aims to include the motion history of the robot in the steering prediction model. The improved model showed its ability to predict steering commands with high performance compared to the expert operator. However, this model imposed some limitations which will be further discussed in this remaining parts of this thesis.

Contrôle

Browse

Filters

Settings

Sort By

Results per page

Search Results