|

Low Speed Automated Shuttle: From Test Track to on-Road Pilot

Author(s): Charlebois, Dube, Christenson

Poster Presentation:

3D-Charlebois

Abstract:

Background:

In the fall of 2020, TC launched a project to test an Easymile EZ10 low speed automated shuttle. The project was undertaken in collaboration with Invest Ottawa and Easymile. The project began with track testing of the LSAS which led to a 2 week trial at Tunney’s Pasture in Ottawa. A track testing plan to evaluate the reaction of the shuttle to cyclists and pedestrians in a controlled environment was conducted at Area X.O. The trial served as a benchmark to better understand battery usage, safe operation and limitations of advanced automated systems.

Aims:

The objectives were to understand the shuttle’s automated driving capabilities (including safe interactions with pedestrians, cyclists and low speed motor traffic), human factors considerations, battery efficiency, and viability as a first-/last-mile transit solution. Results will inform TC safety approaches for LSASs to support the development of guidance documents and TC collaboration on international standards.

Methods:

The vehicle was assessed at Area X.O using an adult, cyclist and child pedestrian dummies going through scenarios that would resemble city-like interactions such as intersection crossing. Care was taken in identifying real world threats that the shuttle could face during its 2-week deployment on the road. The road trial began by mapping the site and noting possible issues that could pose a risk during the trial. Steps were then taken to mitigate or remove these risks from the trial. The shuttle operated at a maximum speed of 15 km/h on a 1.5 km loop servicing 4 predefined stops on the Tunney’s Pasture campus in Ottawa. Extra precautions were taken into consideration with Covid19 safety measures in place.

Results:

The track testing demonstrated that the shuttle could mitigate collisions in 17 different test scenarios involving left turns, right turns and take-over scenarios. No impact was recorded during the track testing where edge cases were studied.
The 2-week road trial took place in various weather conditions which provided challenges such as snow covered ground and blowing leaves, both of which affected the operation of the automated shuttle. Some scenarios prepared for track testing were encountered during the on-road pilot, such as cyclists cutting off the shuttle and Canada geese crossing its path.

Discussion:

The information gathered during the road trial was very useful as it included data on uncontrollable parameters such as vehicles that park outside of designated parking spots and encroach on the shuttle’s path. The EasyMile automated shuttle was designed to slow down when driving next to pedestrians and cyclists to reduce safety risks. After conducting the on-road portion of the pilot it was found that many of the scenarios tested on the track came to fruition in the real world, meaning the methodology used to choose those scenarios was successful. The shuttle had some difficulty operating without delay when there were a lot of blowing leaves which led to questions around how it would handle snow, or heavy rain.

Conclusions:

More adverse environmental effects could be explored in future trials.