This content is not included in
your SAE MOBILUS subscription, or you are not logged in.
Optimizing Maneuver Length for Autonomous Obstacle Avoidance Maneuver with Considerations for Controllability and Passenger Comfort on Low Friction Surfaces
Journal Article
12-05-02-0014
ISSN: 2574-0741, e-ISSN: 2574-075X
Sector:
Topic:
Citation:
Spike, N., Chopp, D., Kurup, A., Bos, J. et al., "Optimizing Maneuver Length for Autonomous Obstacle Avoidance Maneuver with Considerations for Controllability and Passenger Comfort on Low Friction Surfaces," SAE Intl. J CAV 5(2):165-175, 2022, https://doi.org/10.4271/12-05-02-0014.
Language:
English
Abstract:
In order for autonomous vehicles to be widely adopted, they must be able to
operate in all conditions possible in the regions they are operating. In
northern climates, this means they must be able to operate on low friction due
to the presence of ice or snow. An autonomous vehicle performing an obstacle
avoidance maneuver in the form of a double-lane change maneuver must account for
low friction when present to ensure a safe and comfortable ride for passengers.
This work presents a graphical optimization method for determining a minimum
maneuver distance based on surface friction coefficient, which is constrained by
cross-track error and lateral acceleration. The optimization has been performed
in simulation and refined with hardware results from an autonomous test vehicle,
showing the optimal maneuver length being dominated by a constraint on lateral
acceleration on surfaces with a friction coefficient close to that of packed
snow and above and by the cross-track error constraint on icy surfaces. When the
friction coefficient is known precisely, a lookup table based on friction
coefficient can be used to determine the maneuver length. If the friction
coefficient is estimated from broad categories of surface type such as ice,
snow, or pavement, the maneuver length is fixed at 120 m for icy surfaces
(assuming the ice is the lowest friction encountered by the vehicle during
hardware testing) and 70 m for all other surfaces. This work shows the process
of optimizing maneuver length across friction surfaces for a single velocity,
which would be repeated for the range of velocities a production autonomous
vehicle could be assumed to operate during normal use.