DAY 1: Tutorials

CHAMP is an open source development framework for building new quadrupedal robots and developing new control algorithms. The default locomotion controller is based on “Hierarchical controller for highly dynamic locomotion utilizing pattern modulation and impedance control : implementation on the MIT Cheetah robot”. In this workshop, Juan would be demonstrating the software tools and libraries available in the CHAMP framework.

Perception and planning are fundamental elements of autonomous mobility. In this tutorial we focus on the planning aspect. Dynamic constraints of an autonomous vehicle and comfort of passengers (if any) are important constraints in designing its path or trajectory. Accordingly, smooth variation in path curvature together with low computational complexity in generating it, pose the basic problem in path planning. We discuss a path planning method based on the Four Parameter Logistic Curve (4-PL), an explicit assay curve with its origins in ecology and medicine. Analytical results are deduced for satisfying obstacle avoidance, maximum curvature, end-curvature, and heading angle constraints using the 4-PL curve. In the next part, we discuss the problem of following a pre-defined path as a tracking control problem by considering a virtual target-based guidance approach. A linear analysis of the resulting cross-track error is discussed. Last part of the tutorial looks at a diverse trajectory design approach based on bifurcation theory. Towards the end, we emphasize the inter-linkages between trajectory design and a higher state of autonomy.

In this presentation, we will first propose an introduction to the Human-Robot Interaction (HRI) domain at large. We will highlight the specificities of the domain compared to robotics, i.e., what it means to take into account the human in the loop, based on models issued from theories from psychology and philosophy. Then, we will continue with a presentation of the LAAS HRI architecture (DACOBOT) with a focus on: navigation, task planning, decision-making, knowledge management and motion planning. We will finish by a presentation of an instantiation of this architecture through a use case explanation and associated video.