Goal Blending for Responsive Shared Autonomy in a Navigating Vehicle

Authors: Yu-Sian Jiang, Garrett Warnell, Peter Stone5939-5947

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We validate the proposed approach by performing a human study involving an intelligent wheelchair and compare GBSA to a representative servo-level shared control system that uses a policyblending approach. The results of both quantitative performance analysis and a subjective survey show that GBSA exhibits significantly better system responsiveness and induces higher user satisfaction than the existing approach.
Researcher Affiliation Collaboration Yu-Sian Jiang,1 Garrett Warnell, 2 Peter Stone 1,3 1 The University of Texas at Austin 2 Army Research Laboratory 3 Sony AI
Pseudocode Yes Algorithm 1 Overall Flow of GBSA. ... Algorithm 2 Blended Goal Optimization.
Open Source Code No No explicit statement or link providing access to source code for the methodology described in this paper was found.
Open Datasets No The paper describes using a 'simulated wheelchair' and 'local costmap based on the robot s sensor data' but does not specify a publicly available dataset or provide access information for the data collected from the simulation.
Dataset Splits No The paper describes a human study with participants but does not provide specific training, validation, or test dataset splits in terms of percentages or sample counts for data partitioning.
Hardware Specification No The paper mentions a 'simulated SICK LMS111 laser scanner' and 'simulated wheelchair' for the experimental setup, but does not specify the actual hardware (CPU, GPU models, or computing resources) used to run these simulations or experiments.
Software Dependencies No The paper mentions implementation in 'ROS' but does not provide specific version numbers for ROS or any other key software components, libraries, or solvers.
Experiment Setup Yes where ρ is a pre-specified parameter that denotes the distance at which local goals are to be placed (we use ρ = 2.0m in our experiments). ... where both vi,t and vr,t are normalized to unit vectors for the computation above, k is a user-specified scalar that determines the relative strength between the inertial force and the repulsion force, and C(x) is the local costmap value at position x. We set k = 1.1 based on our desire for blended goals to be located in front of obstacles, which will happen if the repulsion force from the obstacle is weighted larger than the user-goal inertial force. ... where w1 and w2 are specified coefficients for the two deviation terms (empirically chosen as w1 = 1 and w2 = 10 in our experiments).