Alain Beaulieu

Dynamic encirclement of a moving target using decentralized nonlinear Model Predictive Control

Dynamic encirclement is a tactic which can be employed by a group of UAVs to neutralize a target by restricting its movement, or provide constant surveillance of a target. The aim of the UAVs in the formation is to move into a position close to the target and establish a moving formation around the target. In this paper, the problem of creating a dynamic circular formation around a moving target is considered, and a Decentralized Model Predictive Control (DMPC) policy is formulated. Using theoretical results, a stabilizing control policy is derived, and the policy is validated through simulation results. Furthermore, we examine the effects of communications between the UAVs and the use of a model target on the performance of the UAVs. The contributions of this paper are the extension of the dynamic encirclement tactic to the case of a group of UAVs and a moving target, the consideration of a target model and communications, and the application of theoretical stability analysis to the problem.

Encirclement of multiple targets using model predictive control

Two teams of Unmanned Aerial Vehicles (UAVs) are used in the encirclement of two targets at the same time. Encirclement is defined as the situation in which a target is isolated and surrounded by a group of UAVs. It is a tactic that can be employed by a team of UAVs to neutralize a target by restricting its movement due to a containment motion near the target while maintaining a formation around it. In this paper, the problem of choosing the correct target to create a dynamic circular formation is considered and a Decentralized Model Predictive Control (DMPC) policy is formulated. From simulation results the derived Model Predictive Control (MPC) policy is effective for the case of two teams of UAVs encircling two stationary targets, and two teams of UAVs encircling two moving targets. The contributions of this paper are the application of MPC to the problem of encirclement, the explicit objective of a dynamic circular formation around the target, and the ability of each team to choose its correct target.

Using multiple Quadrotor aircraft and Linear Model Predictive Control for the encirclement of a target

A Multi-Unmanned Aerial Vehicle (UAV) team formed from two or more UAVs is used in the encirclement of a target. Encirclement is defined as the situation in which a target is isolated and surrounded by a UAV team in order to maintain awareness and containment of that target. In this paper, the problem of maintaining a circular path around a target is considered and a Linear Model Predictive Control (LMPC) strategy is implemented on a team of Qball-X4 quadrotor aircraft in order to follow the circular path. The linear plant controlled by the LMPC is a combination of process models found through system identification and a linear cartesian to polar transformation. A collision avoidance system, based on potential fields, is successfully implemented between the Qball-X4 quadrotors. The contribution of this paper lay in the application of LMPC to the problem of encirclement using a team of Qball-X4 quadrotors and the ability of these UAVs to apply a collision avoidance policy.

Unmanned Aerial Vehicle formation flying using Linear Model Predictive Control

A team of three Unmanned Aerial Vehicles (UAVs) accomplishes a line abreast, triangular and cross formation based on high-level Linear Model Predictive Control (LMPC). All flight tests respect Reynolds rules of flocking, where the UAVs avoid collisions with nearby flockmates, attempt to match velocity of other team members and attempt to stay close to other flockmates. A linear system identification model is at the base of the error dynamics describing the formation control algorithm. The main contribution of this paper lies in the use of LMPC to implement multiple formations on UAVs in simulation and using the Qball-X4 quadrotor.

Verifying Protocol Conformance Using Software Model Checking for the Model-Driven Development of Embedded Systems

To facilitate modular development, the use of state machines has been proposed to specify the protocol (i.e., the sequence of messages) that each port of a component can engage in. The protocol conformance checking problem consists of determining whether the actual behavior of a component conforms to the protocol specifications on its ports. In this paper, we consider this problem in the context of the model-driven development (MDD) of embedded systems based on UML 2, in which UML 2 state machines are used to specify component behavior. We provide a definition of conformance which slightly extends those found in the literature and reduce the conformance check to a state space exploration. We describe a tool implementing the approach using the Java PathFinder software model checker and the MDD tool IBM Rational RoseRT, discuss its application to three case studies, and show how the tool repeatedly allowed us to find unexpected conformance errors with encouraging performance. We conclude that the approach is promising for supporting the modular development of embedded components in the context of industrial applications of MDD.

Using Linear Model Predictive Control via Feedback Linearization for dynamic encirclement

An Unmanned Aerial Vehicle (UAV) team formed from two or more UAVs is used in the autonomous encirclement of a stationary target in simulation. The encirclement tactic is defined as the situation in which a target is surrounded by a UAV team in formation. This tactic can be employed by a team of UAVs to neutralize a target by restricting its movement. A combination of Linear Model Predictive Control (LMPC) and Feedback Linearization (FL) is implemented on a team of UAVs in order to accomplish dynamic encirclement. The linear plant, representing each UAV, is found through System Identification then linearized using an FL technique. The contributions of this paper lay in the application of LMPC and FL to the problem of encirclement using an autonomous team of UAVs in simulation.

Merging of octree based 3D occupancy grid maps

A technique for merging 3D octree based occupancy grid maps is proposed and implemented. Octrees are a memory efficient way to represent a 3D environment by recursively subdividing space at multiple depths in a tree structure. The use of of an octree representation of a 3D environment allows large environments to be mapped while limiting the amount of memory used in comparison to other techniques. When multiple robots are used to map an environment a more accurate map of a larger space can be produced in less time. In this paper, the problem of merging octree based occupancy grid maps from independent robots into one global map of their environment is explored. Techniques are introduced to address information from sources coming from multiple depths in the map as well as relative transformations between maps that are not axis aligned. These techniques allow the octree representation of an environment to be extended to multiple robots. The application of these techniques is demonstrated by merging maps built by robots in a simulated environment. The contribution of this work lies in the introduction of a feasible method of merging memory efficient maps of a 3D environment. The results obtained in this paper demonstrate that the proposed strategies for octree based map mergers are valid.

Verifying UML-RT protocol conformance using model checking

In UML-RT, capsules communicate via protocols which connect capsule ports. Protocol StateMachines (PSMs) allow the description of the legal message sequences of a port and are potentially very useful for the modular development and verification of systems. However, it is unclear how exactly conformance of a capsule to its PSMs should be defined and how this can be checked automatically. In this paper, we provide a definition of protocol conformance and show how software model checking can be used to check protocol conformance automatically. We describe the design and implementation of a tool that checks the conformance of a capsule with Java action code with respect to the PSMs of all its ports. The results of the validation of the tool on three case studies are summarized.

UAVs in Formation and Dynamic Encirclement Via Model Predictive Control

Abstract Switching between the formation flight tactic and the dynamic encirclement tactic for a team of Unmanned Aerial Vehicles (UAVs) is done using a decentralized approach. A team formed from N UAVs, accomplishes a line-of-breast formation then dynamic encirclement around a desired target. A high-level Linear Model Predictive Control (LMPC) policy is used to control the UAV team during the execution of the required formation tactic, while a combination of decentralized LMPC and Feedback Linearization (FL) is implemented on the UAV team to accomplish dynamic encirclement. During the simulations, Reynolds rules of flocking are respected. The linear plant, representing each UAV, is found through System Identification. The main contribution of this paper lies in the use of LMPC to implement multiple UAV tactics while ensuring stability and robustness of the system during tactic switching.

Linear Model Predictive Control for the encirclement of a target using a quadrotor aircraft

Encirclement is a task accomplished by an Unmanned Aerial Vehicle (UAV) in order to maintain awareness and containment of a given target. The aim of the UAV encircling this target is to maintain close proximity at all times. In this paper, the problem of maintaining a circular path around a target is considered and a Linear Model Predictive Control (LMPC) strategy is implemented on a Qball-X4 quadrotor aircraft in order to follow the circular path. A linear model for the two-dimensional movement of the UAV and its respective MP controller has been designed in MATLAB Simulink, simulated in a X-Plane/MATLAB interface and implemented on the actual vehicle in real-time. The results of the LMPC in simulation are compared to those found while implementing the algorithm on a physical platform. The contributions of this paper lay in the implementation of an autonomous Linear MP controller for the encirclement of a stationary target by a Qball-X4 quadrotor.