Periodic Reporting for period 2 - CREME (Control of light vehicle-manipulator systems)
Période du rapport: 2023-02-01 au 2024-07-31
In this project, we aim to develop a control framework that enables light and freely moving autonomous robots that are able to perform complex and versatile operations, involving highly precise and forceful physical interactions, in challenging environments.
Specifically, the robots should not only be able to observe and monitor but should also be able to interact physically with their environment. Moreover, the robots should be multi-purpose robots, which are not only designed to perform one particular task, like for instance a lawn mower robot, but should instead be equipped with manipulator arms and be able to perform a wide variety of operations with highly precise manipulating forces.
Freely moving robots with manipulator arms that give them interaction capabilities are denoted vehicle-manipulator systems, and making them autonomous, light and cheap, sparks exciting perspectives in marine, domestic, logistic, aerial and space contexts.
We will specifically target the underwater environment, as this poses particularly severe challenges including unmodelled time-varying hydrodynamic forces, uncertain hydrodynamic coefficients and strong environmental disturbances.
Specifically, the robots should not only be able to observe and monitor but should also be able to interact physically with their environment. Moreover, the robots should be multi-purpose robots, which are not only designed to perform one particular task, like for instance a lawn mower robot, but should instead be equipped with manipulator arms and be able to perform a wide variety of operations with highly precise manipulating forces.
Freely moving robots with manipulator arms that give them interaction capabilities are denoted vehicle-manipulator systems, and making them autonomous, light and cheap, sparks exciting perspectives in marine, domestic, logistic, aerial and space contexts.
We will specifically target the underwater environment, as this poses particularly severe challenges including unmodelled time-varying hydrodynamic forces, uncertain hydrodynamic coefficients and strong environmental disturbances.
Control systems for vehicle-manipulator systems (VMSs) must handle the effects of complex, unmodelled dynamics and disturbances affecting the VMS. These effects are particularly severe in underwater environments, including highly complex hydrodynamics and strong ocean currents. We have thus developed novel control systems that make the VMS robust to such perturbations, i.e. their effect on the motion of the VMS is strongly reduced. We have used mathematical analysis to develop and prove the stability and performance of the proposed control algorithms and have validated the theoretical results through numerical simulations and full-scale experiments.
VMSs with a small and light base compared to the size of the manipulator arm it carries are denoted light-VMSs. There will always be coupling effects between the motion of the manipulator arm and the base of the vehicle, but for VMSs with a large, heavy base, like a work-class ROV, the motion of the arm will only affect the motion of the base slightly. For light-VMSs, on the other hand, there will be a strong coupling. Moreover, all VMSs are redundant systems, as they can always choose between moving their base and arm to position their end-effector. Therefore, we have developed control methods that handle this redundancy and utilize it to achieve several tasks simultaneously while also considering the coupling effects. Task-based methods provide a structured way of building up autonomy in a system, where many tasks can be defined with different priorities attached to them, and the algorithm plans and executes the pertinent VMS motion online. Our methods ensure strict priority between the different tasks, such that low-priority tasks do not impact the execution of high-priority tasks like collision avoidance and other strict safety requirements.
To achieve true autonomy for underwater vehicles, energy autonomy must be achieved. Today, underwater vehicles rely on subsea infrastructure with docking stations or a surface supply vessel to recharge their batteries. It will be revolutionary for our ability to explore, monitor, and care for our oceans if we can develop underwater vehicles that can extract the energy they need from the ocean. In this project, we have investigated whether articulated underwater VMSs, called Articulated Intervention-AUVs (AIAUVs), can harvest energy from the wakes downstream of bluff bodies when currents pass by. We have obtained promising results, showing that energy can be harvested and that there exists a clear optimal position in which the most energy is generated. We have, furthermore, developed control algorithms that enable the AIAUV to locate and stay in this optimal position for energy harvesting.
VMSs with a small and light base compared to the size of the manipulator arm it carries are denoted light-VMSs. There will always be coupling effects between the motion of the manipulator arm and the base of the vehicle, but for VMSs with a large, heavy base, like a work-class ROV, the motion of the arm will only affect the motion of the base slightly. For light-VMSs, on the other hand, there will be a strong coupling. Moreover, all VMSs are redundant systems, as they can always choose between moving their base and arm to position their end-effector. Therefore, we have developed control methods that handle this redundancy and utilize it to achieve several tasks simultaneously while also considering the coupling effects. Task-based methods provide a structured way of building up autonomy in a system, where many tasks can be defined with different priorities attached to them, and the algorithm plans and executes the pertinent VMS motion online. Our methods ensure strict priority between the different tasks, such that low-priority tasks do not impact the execution of high-priority tasks like collision avoidance and other strict safety requirements.
To achieve true autonomy for underwater vehicles, energy autonomy must be achieved. Today, underwater vehicles rely on subsea infrastructure with docking stations or a surface supply vessel to recharge their batteries. It will be revolutionary for our ability to explore, monitor, and care for our oceans if we can develop underwater vehicles that can extract the energy they need from the ocean. In this project, we have investigated whether articulated underwater VMSs, called Articulated Intervention-AUVs (AIAUVs), can harvest energy from the wakes downstream of bluff bodies when currents pass by. We have obtained promising results, showing that energy can be harvested and that there exists a clear optimal position in which the most energy is generated. We have, furthermore, developed control algorithms that enable the AIAUV to locate and stay in this optimal position for energy harvesting.
We have developed new fundamental Lyapunov theory results for uniform practical asymptotic stability. These results apply to a large class of nonlinear dynamical systems, and we have utilized them to solve the position control problem of articulated intervention-AUVs (AIAUVs).
We have also developed the Adaptive Generalized Super-Twisting Algorithm (AGSTA), which provides control with robustness to perturbations like unmodelled dynamics and environmental disturbances for a large class of nonlinear systems, including VMSs both on land, in air, and underwater. Specifically, the proposed control method handles state-dependent disturbances as well as time-dependent ones, handles unknown control coefficients, does not require knowledge of the bounds on the perturbations, and gives good performance in practice through adaptively finding good gains.
We have developed new control methods for robust task-priority motion planning and control of VMSs, which take into account the coupling of kinematics and dynamics that is particularly challenging for light-VMSs and provide robustness to strong environmental disturbances and unmodelled dynamics. The framework ensures strict priority between the tasks, such that the lower-priority tasks cannot interfere with the higher-priority safety tasks, and also includes impedance control for physical interaction operations.
We have initiated a novel line of research to enable underwater VMSs to harvest the energy they need from the oceans, creating the energy autonomy required for truly autonomous UVMSs. Through mathematical analysis, high-fidelity simulations, and experiments, we have shown that the energy can be harvested by the AIAUV in vortex wakes, that there exists a clear optimal position in which the most energy is generated, and we have developed control algorithms that enable the AIAUV to locate and stabilize this position. We expect to further increase the efficiency of the energy harvesting and validate the theoretical findings in full-scale experiments.
We have also developed the Adaptive Generalized Super-Twisting Algorithm (AGSTA), which provides control with robustness to perturbations like unmodelled dynamics and environmental disturbances for a large class of nonlinear systems, including VMSs both on land, in air, and underwater. Specifically, the proposed control method handles state-dependent disturbances as well as time-dependent ones, handles unknown control coefficients, does not require knowledge of the bounds on the perturbations, and gives good performance in practice through adaptively finding good gains.
We have developed new control methods for robust task-priority motion planning and control of VMSs, which take into account the coupling of kinematics and dynamics that is particularly challenging for light-VMSs and provide robustness to strong environmental disturbances and unmodelled dynamics. The framework ensures strict priority between the tasks, such that the lower-priority tasks cannot interfere with the higher-priority safety tasks, and also includes impedance control for physical interaction operations.
We have initiated a novel line of research to enable underwater VMSs to harvest the energy they need from the oceans, creating the energy autonomy required for truly autonomous UVMSs. Through mathematical analysis, high-fidelity simulations, and experiments, we have shown that the energy can be harvested by the AIAUV in vortex wakes, that there exists a clear optimal position in which the most energy is generated, and we have developed control algorithms that enable the AIAUV to locate and stabilize this position. We expect to further increase the efficiency of the energy harvesting and validate the theoretical findings in full-scale experiments.