TU Darmstadt / ULB / TUbiblio

Linear Model-Predictive Control of Cooperative Multi-Vehicle Systems for Time-Dependent Sampling Applications

Zeng, Dequan (2011):
Linear Model-Predictive Control of Cooperative Multi-Vehicle Systems for Time-Dependent Sampling Applications.
Darmstadt, Technische Universität Darmstadt, Department of Computer Science & Department of Mathematics, [Diploma Thesis or Magisterarbeit]

Abstract

Employing autonomous multi-vehicle systems to perform tasks like surveillance or environment exploration has many advantages, e.g. they reduce the risk of life-threatening missions that might otherwise be carried out by human-piloted vehicles. For such systems, one of the key problems is the optimal control of cooperative mobility. In this thesis, an approach concerning this problem is developed and investigated, which can be employed to time-dependent sampling applications. The optimal control problem of cooperative multi-vehicle systems investigated in this thesis cov- ers two specific aspects: task allocation by priority and obstacle avoidance. The system model is built upon discrete logical rules and continuous vehicle dynamics. Hence, the problem constraints together with the objective function form a hybrid optimal control problem. Approximating this problem by Mixed-Integer Linear Program (MILP) reduces the complexity of the problem and improves the computational efficiency. Solutions of the hybrid problem are obtained by a model-predictive control (MPC) approach, since its feedback mechanism allows the multi-vehicle system to react to a changing environment. An existing decentralized MILP-based MPC approach KRvS11 is extended to cover extra prob- lem constraints with respect to the time-dependent task allocation and obstacle avoidance. The optimal control problem is modeled employing the mixed-logical-dynamical (MLD) framework. In this thesis, the validity of the developed method will be proved, its applicability to real-time (small scale) systems will also be shown to be convincing. The presented investigations set up a possible basis for further research and development of optimal control of cooperative multi-vehicle systems.

Item Type: Diploma Thesis or Magisterarbeit
Erschienen: 2011
Creators: Zeng, Dequan
Title: Linear Model-Predictive Control of Cooperative Multi-Vehicle Systems for Time-Dependent Sampling Applications
Language: English
Abstract:

Employing autonomous multi-vehicle systems to perform tasks like surveillance or environment exploration has many advantages, e.g. they reduce the risk of life-threatening missions that might otherwise be carried out by human-piloted vehicles. For such systems, one of the key problems is the optimal control of cooperative mobility. In this thesis, an approach concerning this problem is developed and investigated, which can be employed to time-dependent sampling applications. The optimal control problem of cooperative multi-vehicle systems investigated in this thesis cov- ers two specific aspects: task allocation by priority and obstacle avoidance. The system model is built upon discrete logical rules and continuous vehicle dynamics. Hence, the problem constraints together with the objective function form a hybrid optimal control problem. Approximating this problem by Mixed-Integer Linear Program (MILP) reduces the complexity of the problem and improves the computational efficiency. Solutions of the hybrid problem are obtained by a model-predictive control (MPC) approach, since its feedback mechanism allows the multi-vehicle system to react to a changing environment. An existing decentralized MILP-based MPC approach KRvS11 is extended to cover extra prob- lem constraints with respect to the time-dependent task allocation and obstacle avoidance. The optimal control problem is modeled employing the mixed-logical-dynamical (MLD) framework. In this thesis, the validity of the developed method will be proved, its applicability to real-time (small scale) systems will also be shown to be convincing. The presented investigations set up a possible basis for further research and development of optimal control of cooperative multi-vehicle systems.

Place of Publication: Darmstadt
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Simulation, Systems Optimization and Robotics Group
04 Department of Mathematics
04 Department of Mathematics > Optimization
Date Deposited: 10 Jul 2019 08:52
Export:
Suche nach Titel in: TUfind oder in Google

Optionen (nur für Redakteure)

View Item View Item