General Motors Almost Doubles Super Cruise Coverage, but Unscalable Mapping Approaches Present a Looming Risk

General Motors (GM) began 2024 with news breaking of the end of its Ultra Cruise program, which had been announced in 2021 as the next generation of its autonomous driving solution, 3and which was intended to deliver automation of 95% of driving tasks on 2 million roads. This system was set to leverage Qualcomm’s Snapdragon Ride System-on-Chip (SoC), as well as camera/radar/Light Detection and Ranging (LiDAR) sensor fusion to deliver a significant leap forward beyond the capabilities of the Super Cruise system that GM first launched in 2017, automating more driver functions, and greatly expanding the operational design domain into more road types, including the urban environment.

Now, GM is shifting toward a more evolutionary approach, announcing in February 2024 the expansion of its Super Cruise system to more models, including models such as the Traverse and Acadia from GM’s mass market brands, with operation across 750,000 miles of roads in North America. Since its introduction in 2017, Super Cruise has delivered more than 160 million miles of hands-free operation, with GM reporting no accidents so far. This represents growth from the 34 million miles of hands-free operation that GM announced in August 2022. Therefore, in the space of less than 18 months, GM was able to almost double the number of miles of roads that can be addressed with hands-free driving, while use of Super Cruise has grown at a significant rate.

Therefore, GM’s focus is now firmly directed toward hands-free highway driving, and expanding the number of maneuvers, vehicles, and roads encompassed by its long-standing Super Cruise product. As more Original Equipment Manufacturers (OEMs) face tough decisions regarding their own autonomous vehicle investments, an industry-wide pivot toward hands-free highway driving is the likely next step for passenger vehicle automation.

Within hands-free highway vehicle automation, there are several avenues for differentiation between OEMs:

• Coverage: This will include not only the total number of miles of road that can be supported, but also the extent to which extremes of road curvature, road marking quality, and weather/lighting can be accommodated. GM claims coverage 6X larger than the next biggest hands-free system. The factors that dictate which roads can be supported by a Level 2+ system will vary by market, but the current major binding constraint is the extent of the High-Definition (HD) map layer.
• Maneuvers: Within Level 2+, there is considerable scope for which driving tasks or maneuvers can be performed hands-free on the driver’s behalf. For example, automatic overtaking, automatic lane merging, or, in GM’s case, automatic trailering.
• Human-Machine Interface (HMI): Hands-free, but supervised driving is a novel concept to most consumers, and while valued by early adopters (GM reports 80% of drivers claiming increased relaxation), widespread adoption will require a robust and natural HMI to foster consumer trust. BMW has demonstrated a concept whereby a driver’s glances toward rearview mirrors will trigger lane change maneuvers, giving the driver more ownership of monitoring the driving environment.
• Over-the-Air (OTA) Updates and Lifecycle Management: Both the coverage and maneuvers supported by the Level 2+ system can evolve over time with the correct OTA channel and lifecycle management strategy. Indeed, many existing Super Cruise customers will receive incremental updates through 2025.

In GM’s Super Cruise system, as well as many other Level 2+ systems, digital maps will be a key factor dictating the coverage of the Level 2+ experience, which in turn will be a key differentiator between competing Level 2+ systems. Currently, GM uses LiDAR-based maps supplied by Dynamic Mapping Platform (DMP), which acquired GM’s original HD mapping supplier, Ushr. DMP uses dedicated survey vehicles equipped with LiDAR to curate a survey-grade map, which is critical for operating GM’s Super Cruise.

Since the initial rollout of Super Cruise, the extent of supported roads has grown at an accelerated rate, first growing from 131,000 miles to 400,000 miles between 2018 and 2022 (269,000 miles in 4 years), and now growing by 350,000 miles in less than 2 years. This acceleration can be attributed to greater automation of the map creation process, which formerly had a number of highly manual bottlenecks. Another approach to stretching the viability of the dedicated mapping fleet is using the detection of deltas between ground truth and the map by vehicles in the field to point scarce mapping vehicles to areas that need updating.

However, the dedicated mapping fleet approach is not scalable. These dedicated vehicles are expensive, costing between US$125,000 and US$200,000. It is simply impossible to build a fleet that can map the entire world. Even as the mapped area grows, more of the fleet’s resources must be used to simply maintain what is already promised to the consumer. While increased automation and intelligent deployment of mapping vehicles can improve scalability, any system with coverage that is dependent on a dedicated mapping fleet is doomed to fail.

Ultimately, the only fleet large enough to map the world’s roadways is the fleet of millions of passenger vehicles that use these roadways every day. Crowdsourcing information from connected and sensor-equipped passenger vehicles presents a clear path toward scaling HD maps, as well as maintaining these maps with a fast time to reflect reality.