-->
Robex est une des équipes du Lab-STICC. Elle fait partie du pole IA&O (Intelligence Artificielle et Océan).
Tous les ans sont organisés les Robex days.
Pendant ces journées, il y a des exposés internes à Robex.
Comme tous les ans, cette rencontre se tiendra à l'ENSTA Bretagne :
Swim with brackets
In this presentation, we show that skating or swimming corresponds more or less to a control based on Lie brackets. For this purpose, we consider a snake like car, standing on five ice skates, which is designed to move on a frozen lake. We propose to apply the lie brackets between the two input fields. It will be shown that this creates a new direction in the state space that can now be controlled. Another contribution of this works is to show that it is possible to move extremely fast along a direction created by Lie brackets.
Interval guaranteed matrix decompositions
In this talk, we consider the design of guaranteed interval matrix computations, such as matrix inversion, LU decomposition, Cholesky decomposition and QR decomposition from non-guaranteed floating-point approximations. We also the discuss the application to guaranteed linear programming.
Stable Cycles for autonomous navigation
Navigating GNSS-denied environments presents significant challenges, particularly for underwater robots. This work proposes a novel approach to autonomous navigation that leverages stable cycles to eliminate the need for localization. Using prior environmental knowledge, such as bathymetric maps, the robot follows a predefined trajectory governed by a timed automaton. By dynamically adjusting the automaton’s timing based on limited bathymetric measurements from an echosounder, the robot stabilizes its trajectory along isobaths. The first key contribution of this method is achieving robust trajectory stabilization along isobaths by precisely controlling the cycle’s position. The second contribution enables seamless transitions between stable cycles. The robot navigates in dead-reckoning to reach the capture basin of the next cycle, where a new automaton takes over to stabilize its trajectory. By taking advantage of the link between stable cycles, we can determine worlds of cycles linked together by a reachability relationships, and navigate from world to world to explore an area. This innovative approach offers a reliable solution for navigating in complex underwater environments without GNSS, ensuring the robot can survey an environment without getting lost.
Ellipsoid tether model for collision avoidance in a fleet of ROVs
During collision avoidance, the tether of Remote Operated Vehicle (ROV) is subject to entanglement with obstacles or other ROVs’tether. This specificity means that the classic literature on multi-robot obstacles avoidance is not suitable to tethered multi-robot scenarios. Thus, we propose a guarantee ellipsoid model of the ROV’s tether and its obstacles to perform a collision avoidance method for a fleet of ROVs, low in calculation. This model guarantees that if an obstacle is not inside or partially inside the ellipsoid model, then we are sure that the tether is not collide with that obstacle. This model is simple and can provide a guaranteed proof of non-collision without any knowledge of the tether shape or its dynamic, only its two attachments point and its length. A collision avoidance strategy has been developed based on potential field methods, tether’s length management, and bypass path. When several ROVs are involved, personalities are added to ROV to obtain different behaviors in the same configuration and so limit the case of minimal local during collision avoidance. Simulations show the effectiveness of the method with several scenarios and the limits of the method are discussed.
Boundary approach to characterize the inner and outer approximation of the image of a disk
Calculating directly the inner and outer approximation of the image of a set by a function can be challenging. Then, it is sometimes prefered to compute the image of the boundary of the set instead. However, boundary- based methods are subject to the apparition of fake boundaries in the image set. As they add pessimism when characterizing the inner approximation of the image set, this paper introduces the notion of Box Chains to simplify their detection and their suppression. The characterization of the inner and outer approximation of the image set in the case of a function from the unit disk D to R2 will be considered, with two examples.
Propulsion d'un AUV par palme
Inspirée par le vivant, la palme a été inventé par l'Homme pour nager plus rapidement et efficacement.
Utiliser une palme pour propulser un AUV semble donc être une option qui mérite d'être étudiée. Par rapport à une propulsion par hélice traditionnelle, on peut imaginer certains avantages à vérifier:
- moins perturbant (perturbations sonores, soulèvement de sédiment)
- plus économique (moins d'actionneurs nécessaires)
- plus efficace (mouvements plus naturels, moins de traînée)
PeliKwad: air-water drone deployed at Dronathlon 2024
On 7 and 8 October 2024, the AutoRob (autonomous robots) team represented LabISEN (the ISEN laboratory) at the Dronathlon challenge organised in Toulon by the French Navy. The team took part in 2 of the 10 groups of French manufacturers and laboratories selected by the Navy to demonstrate defence-related capabilities linked to using drones at sea. We will begin by presenting the context of the Dronathlon, the organisation of the project over a period of around a year, in order to give some feedback on this type of challenge and the obstacles encountered during its set-up. We will then move on to the presentation of the concrete demonstrations produced by the laboratory. First, Henrique and Titouan will present the development of a prototype amphibious quadricopter drone (air-water): the PeliKwad. Using rapid prototyping and a simple concept adaptation, Henrique will detail the approach chosen to add this hybrid aspect. Titouan will then develop the electronic and IT architecture designs using the PixHawk system, and their implementation. Finally, special attention will be paid to the mission and results obtained during the demonstration day.