One of the biggest concerns for road logistics operators in the United States is revenue loss due to an inability to operate in inclement conditions for routes across Northern states.
Fortunately, autonomous trucks can transform the entire road logistics industry for the better – and sooner than we might think, thanks to advancements by Embark Trucks (and a bit of help from our comprehensive weather datasets).
As the longest-operating self-driving truck program in the US, Embark Trucks has been a forerunner in autonomous vehicle (AV) technology for trucks since 2016. Their software powers safe, commercially viable autonomous long-haul trucks, trusted by investors, Fortune 500 shippers, and some of the nation’s largest carriers.
Just like self-driving passenger vehicles, autonomous trucks use sensing technologies such as LiDAR (the use of light to determine distance), radar, and optical cameras to collect visual data from the surrounding environment. This data is then combined with maps and algorithms to make decisions while on the road.
Accurate detection and interpretation of road conditions is crucial to proper autonomous vehicle functionality. AVs are trained to successfully adapt to a variety of road conditions, but weather can present challenges for even the most advanced sensing technology. In addition to on-road testing, Embark required accurate and complete historical weather datasets to fully understand the implications of adverse weather conditions on its self-driving solution. With such a crucial business implication at stake, there was no room for compromise on weather data quality.
Granular weather data pulled on location-based intervals frequently throughout the day was a must for Embark. Using an AerisWeather Flex subscription to obtain the data, Embark was able to develop a comprehensive weather model comprised of over eight billion historical weather data points dating back over 10 years for major routes across the United States. This model increases Embark's confidence in its autonomous trucks' ability to assess and react to the impact of snow at a lane level across the US.
The data was also combined with highway requirements - most highways must be cleared within three hours of snow stopping - allowing Embark to estimate when highways will be cleared at any given location.
To determine operational feasibility and demonstrate the instrumentation’s ability to accurately sense the environment, trucks were test-driven in snowy conditions and data from the truck’s sensors was cross-referenced with AerisWeather’s interpolated conditions. Additionally, this testing process enabled Embark to determine the possible business implications of snowy conditions – specifically, how often a truck would be able to run a particular route in the snow based on historical data.
During on-road testing in February 2022, Embark’s trucks traveled on a 60-mile round-trip route on public roads in Montana - between Clinton and Missoula - in varying intensities of winter weather.
The testing and performance review indicated that Embark's software works within tolerance thresholds for safe operation in snowfall rates up to one-sixth of an inch per hour, and with snow accumulation of one inch on the road over 3 hours. The scope of these conditions captures the majority of snowy weather. Embark estimates its software will operate within acceptable shipper delivery windows approximately 90% of the time on delivery runs under such weather conditions. These test results represent a significant milestone in the development of autonomous trucking technology.
Transportation and Logistics
Machine Learning for Autonomous Trucks