Autonomous shuttles and technologies, from today to tomorrow!

Published on January 29, 2021Propulsion QuebecArticle
Autonomous shuttles and technologies, from today to tomorrow!

As part of its Demonstration and Experimentation initiative, Propulsion Québec held the Crossed Perspectives Forum: autonomous shuttles in Quebec on September 9 and 10. Presented as five virtual conferences, its goal was to review the autonomous shuttle projects that have taken place in Quebec and to share the lessons learned from these experiences.

Following the knowledge that emerged from this exchange, Propulsion Québec is offering the Crossed Perspectives article series highlighting the themes discussed during the Forum. This sixth and final article offers an assessment oftechnologies, from today to tomorrow.

Following the autonomous shuttle experiments carried out over the past few years in Quebec, several experts state that they do not provide sufficient safety to be integrated into current transport systems. Let’s look at the safety-related issues and how autonomous shuttles will evolve over the next few years within our transport system.

Theautonomous shuttle experiments that took place in Quebec, including Keolis’s in Candiac and Transdev’s on the territory of the Olympic Park in Montreal, do not guarantee 100% safe operation in their current format, according to Denis Gingras, full professor in the Department of Electrical Engineering and Computer Engineering at the University of Sherbrooke and director of the Vehicular Intelligence Laboratory. Mr. Gingras believes it is currently far too unsafe to replace a human driver with a fully automated shuttle. Several issues that fail to ensure either passenger safety or the safety of road users were raised and are developed below.

A safety issue

The main safety problems identified during autonomous shuttle experiments in Quebec have been grouped into four categories:

  1. Route deviation: geolocation and trajectory planning issues;
  2. Misinterpretation of the environment: perception problems – detection, classification and tracking of (moving) nearby obstacles;
  3. Poor maneuvers: control problems, sudden accelerations/decelerations, unjustified hard braking;
  4. Erratic passenger interfaces: poorly synchronized passenger boarding and disembarking, poor perception of user behavior.

Autonomous vehicles face major challenges to reach a level of safety equivalent to human driving.As Denis Gingras explains, “the human driver is an excellent driver; the probabilities of a serious accident are very rare. Autonomous vehicle technologies do not currently allow us to reach that same level of safety.”

Applying autonomous shuttle technologies to traditional shuttles

Why not use the onboard systems of autonomous shuttles on traditional shuttles, thereby further enhancing human driving safety?That is the suggestion of our vehicular intelligence expert. “Technologies can be used in the short term more intelligently than full autonomy and full automation of driving by using ADAS (Advanced Driver Assistance Systems). I suggest opting for a shuttle with a human driver that can be instrumented to enhance the safety of human driving with, for example, video cameras and sensors to eliminate blind spots.” Denis Gingras even proposes a pilot project with a traditional shuttle where the automated driving system is integrated and, for a determined period, driving data are collected. “This way we accumulate the largest number of driving scenarios possible and see how the onboard system reacts without endangering passengers and people around the shuttle since it does not take control of the shuttle.”

Addressing a mobility need

According to Denis Gingras, we must first assess mobility needs and determine the real relevance of traditional shuttles with human drivers before even thinking about developing fully autonomous shuttles. “If we have a pilot shuttle project to put forward, it would be important to evaluate the real need, the average number of passengers, determine where the urgent needs are, verify and validate that it really provides a sustainable solution,” says Mr. Gingras. “If there are needs, after a few years, and things go well with traditional shuttles with human drivers, then we can consider automating those shuttles in a safe context.”

To listen to (or listen again!) | Webinar on technologies, from today to tomorrow

With:

The speaker, an expert in vehicular intelligence, shares his experience with the audience on the technological limits of autonomous shuttles and his vision of the future of autonomous shuttles.

Speaker:

  • Denis Gingras, full professor, Department of Electrical Engineering and Computer Engineering, University of Sherbrooke and director, LIV-UdeS (Vehicular Intelligence Laboratory)

This forum, organized by Propulsion Québec, was made possible thanks to our major partner theMontreal Urban Innovation Laboratory and to our collaboratorsJalon Montréal andIVÉO.

The autonomous shuttle industry in Quebec

Several Propulsion Québec members are involved, directly or indirectly, in bringing these autonomous shuttle projects to fruition. To learn more about their expertise in the field, we invite you to visit their various websites:

Also read the first articles in the Crossed Perspectives series:Anatomy of an autonomous shuttle: portrait of a technology that opens up a whole new path, The pioneers: three autonomous shuttle projects in Quebec since 2018, Regulation in Quebec for autonomous shuttles: the current situation, Deployment of autonomous shuttles: social acceptability and citizen receptivity, The transition to mobility that accommodates autonomous shuttles |||NEXT|||.

Continue reading on the topic

View all resources

With the financial support of:

Gouvernement du QuébecGouvernement du CanadaCommunauté métropolitaine de MontréalFaskenHydro-QuébecFonds de solidarité FTQ