The core of self-supervised point cloud learning lies in setting up appropriate pretext tasks, to construct a pre-training framework that enables the encoder to perceive 3D objects effectively. In this paper, we integrate two prevalent methods, masked point modeling (MPM) and 3D-to-2D generation, as pretext tasks within a pre-training framework. We leverage the spatial awareness and precise supervision offered by these two methods to address their respective limitations: ambiguous supervision signals and insensitivity to geometric information. Specifically, the proposed framework, abbreviated as PointCG, consists of a Hidden Point Completion (HPC) module and an Arbitrary-view Image Generation (AIG) module. We first capture visible points from arbitrary views as inputs by removing hidden points. Then, HPC extracts representations of the inputs with an encoder and completes the entire shape with a decoder, while AIG is used to generate rendered images based on the visible points’ representations. Extensive experiments demonstrate the superiority of the proposed method over the baselines in various downstream tasks. Our code will be made available upon acceptance.@en

The operation of Unmanned Aerial Vehicles (UAVs) is often subject to state-dependent alterations and unstructured uncertainty factors, such as unmodelled dynamics, environmental weather disturbances, aerodynamics gradients, or changes in inertia and mass due to payloads. While a large number of autopilot solutions have been proposed to operate UAVs, none of these solutions is able to counteract the effects of state-dependent and unstructured uncertainties online by parameter estimation and adaptive control techniques. This work presents a systematic integration of adaptive control into ArduPilot, a popular open-source autopilot suite maintained by a large community of UAV developers. Adaptation features are embedded in the ArduPilot control structure without altering the original architecture, to allow users to use the autopilot suite as usual. Tests show that the proposed adaptive ArduPilot provides consistent improved performance in several uncertain flight conditions. The source code of the proposed adaptive ArduPilot is released at https://github.com/Friend-Peng/Adaptive-ArduPilot-Autopilot.@en

Adaptive integral sliding mode control (AISMC) is an extension of adaptive sliding mode control which is a way to ensure sliding motion while handling system uncertainties. However, conventional AISMC formulations require to different extent a priori knowledge of the system uncertainty: either the upper bound of the uncertainty or of its time derivative are assumed to be bounded a priori, or the uncertainty is assumed to be parametrized by some structure-dependent factorization. This work proposes a variant of AISMC with reduced a priori knowledge of the system uncertainty: it is shown that Euler-Lagrange dynamics typical of sliding mode literature admit a structure-independent parametrization of the system uncertainty. This parametrization is not the result of structural knowledge, but it comes from basic properties of Euler-Lagrange dynamics, valid independently on the structure of the system. The AISMC control method arising from this parametrization is analyzed in the Lyapunov stability framework, and validated in systems with different structures: a surface vessel and an aerial vehicle.

@en

Artificial-delay control is a method in which state and input measurements collected at an immediate past time instant (i.e. artificially delayed) are used to compensate the uncertain dynamics affecting the system at the current time. This work formulates an artificial-delay control method with adaptive gains in the presence of nonlinear (Euler-Lagrange) under-actuation. The appeal of studying Euler-Lagrange dynamics is to capture many robotics applications of practical interest, as demonstrated via stability and robustness analysis and via robotic ship and robotic aerial vehicle test cases.

@en