Argo AI Launches AV Safety Guidelines to Curb Crashes with Bicyclists

Argo AI, a global autonomous vehicle technology platform company, in partnership with the League of American Bicyclists, has developed and launched technical guidelines to ensure safe interactions between autonomous vehicles (AVs) and cyclists.

The new technical guidelines deliver on Argo AI’s commitment to developing a self-driving system that is trusted by cyclists and enhances the safety of the communities in which they operate, according to the company.

In the U.S., traffic fatalities of cyclists increased 5% in 2020 over 2019, according to the National Highway Traffic Safety Administration (NHTSA).

Keeping these vulnerable road users safe has long been a challenge. With AVs entering our roadways, the challenge is even greater.

Teaming up with the League of American Bicyclists, Argo AI has created a safety blueprint designed to reduce the risk of an autonomous vehicle hitting a cyclist. The partners outlined six technical guidelines for the manner in which a self-driving system (SDS) should accurately detect cyclists, predict cyclist behavior, and drive in a consistent way to effectively and safely share the road.

The first concerns perception. An SDS should designate cyclists as a core object representation within its perception system in order to detect cyclists accurately as opposed to pedestrians, for example. By treating cyclists as a distinct class and labeling a diverse set of bicycle imagery, a self-driving system detects cyclists in a variety of positions and orientations, from a variety of viewpoints, and at a variety of speeds.

Secondly, a system should be built to predict typical cyclist behavior. For example, a cyclist may lane split, yield at stop signs, walk a bicycle, or make quick, deliberate lateral movements to avoid obstacles. Understanding of potential cyclist movement patterns is necessary to best predict their intentions and prepare the self-driving vehicle’s actions.

The third guideline aims to incorporate an understanding of cycling infrastructure. A self-driving system should use high definition 3D maps that incorporate details about cycling infrastructure, like where dedicated bike lanes are located, and include all local and state cycling laws to ensure its self-driving system is compliant. Accounting for bike infrastructure enables the SDS to anticipate cyclists and to maintain a safe distance between the self-driving vehicle and the bike lane.

The fourth point says that an SDS should drive in a consistent and understandable manner. Developers should aim for the technology to operate in a naturalistic way so that the intentions of autonomous vehicles are clearly understood by other road users. In the presence of nearby cyclists or when passing or driving behind cyclists, an SDS should target appropriate speeds in accordance with local speed limits, and margins that are equal to or greater than local laws, and only pass a cyclist when it can maintain those margins and speeds for the entire maneuver.

The fifth guideline recommends an SDS prepare for uncertain situations and proactively slow down. A self-driving system should account for uncertainty in cyclists’ intent, direction, and speed — for instance reducing vehicle speed when a cyclist is traveling in the opposite direction of the vehicle in the same lane. When there is uncertainty, the self-driving system should lower the vehicle’s speed and, when possible, increase the margin of distance.

Finally, the new recommendations say there should be a focus on testing of cyclist scenarios. Developers of self-driving technology should be committed to continuous virtual and physical testing of its self-driving system with a specific focus on cyclist safety in all phases of development. This can include both virtual and physical testing.

Some 846 bicyclists in the U.S. lost their lives in 2019 alone, according to NHTSA.