CAHSOR: Competence-Aware High-Speed Off-Road Ground Navigation in SE(3)

TL;DR

The efficacy of the Competence-Aware High-Speed Off-Road navigation approach on a physical ground robot in both autonomous navigation and a human shared-control setup is demonstrated and it is shown that cahsor can efficiently reduce vehicle instability by 62% while only compromising 8.6% average speed with the help ofron.

Abstract

While the workspace of traditional ground vehicles is usually assumed to be in a 2D plane, i.e., SE(2), such an assumption may not hold when they drive at high speeds on unstructured off-road terrain: High-speed sharp turns on high-friction surfaces may lead to vehicle rollover; Turning aggressively on loose gravel or grass may violate the non-holonomic constraint and cause significant lateral sliding; Driving quickly on rugged terrain will produce extensive vibration along the vertical axis. Therefore, most offroad vehicles are currently limited to drive only at low speeds to assure vehicle stability and safety. In this work, we aim at empowering high-speed off-road vehicles with competence awareness in SE(3) so that they can reason about the consequences of taking aggressive maneuvers on different terrain with a 6-DoF forward kinodynamic model. The model is learned from visual and inertial Terrain Representation for Off-road Navigation (TRON) using multimodal, self-supervised vehicle-terrain interactions. We demonstrate the efficacy of our Competence-Aware High-Speed Off-Road (CAHSOR) navigation approach on a physical ground robot in both an autonomous navigation and a human shared-control setup and show that CAHSOR can efficiently reduce vehicle instability by 62% while only compromising 8.6% average speed with the help of TRON.

Figures (6)

• Figure 1: Challenges of High-Speed Off-Road Ground Navigation in $\mathbb{SE}(3)$.
• Figure 2: Overview of the Competence-Aware High-Speed Off-Road (cahsor) Ground Navigation (Side and Top View) based on Vision ($\psi^{V}$), Inertia ($\psi^I$), and Speed($\psi^{S}$) Representation.
• Figure 3: tron (Left) and Downstream Kinodynamics Learning (Right) Architecture: Flame and temperature denote training and frozen parameters respectively.
• Figure 4: Downstream Kinodynamic Model Prediction with tron and sterling Pretraining Compared to Ground Truth: tron produces both qualitatively and quantitatively accurate predictions, while sterling fails to qualitatively capture the changes in roll and sliding, but only gets the trend of bumpiness.
• Figure 5: Examples of Human-cahsor Shared Autonomy on Rocks, Grass, and Pavement to Limit Bumpiness, Sliding, and Roll.