This week, Zoox revealed the first look of its driving, electric, autonomous vehicle built for the rider, not the driver. We’re happy to confirm that FLIR Systems will supply Zoox with thermal imaging cameras, and we wanted to share more details on our involvement with Zoox and why this is such an important milestone for autonomous vehicles.
As a sensing technology that performs the complex task of detecting and classifying roadway objects in challenging lighting conditions, thermal cameras add additional layers of safety through increased situational awareness and redundancy as part of complex sensor suites, ranging from radar to LIDAR to visible cameras.
The Zoox robotaxi vehicle features FLIR thermal cameras in its bi-directional vehicle, bringing improved visibility to more confidently identify and classify pedestrians, cyclists, and animals in both day and night driving conditions on urban streets, all around the vehicle.
Although thermal imaging is a more recent sensing modality in autonomy, mobility innovators, like Zoox, have come to realize its benefits for improving safety. Thermal imaging is a passive sensing modality that does not require illumination and performs equally well in bright sun glare, headlight glare, and complete darkness, while offering additional awareness in adverse weather conditions such as fog, smoke, rain and snow. In short, by using thermal imaging, ADAS vehicles can recognize and adapt to these changing conditions quickly, and as a result, improve the safety across all levels of autonomy, including automatic emergency braking (AEB). This level of safety is crucial for all occupants and other road users, especially pedestrians and bicyclists that can be more difficult to spot in crowded urban environments, to help avoid injury or potential fatalities.
A History of Thermal Imaging and Automotive Innovation
FLIR’s initial foray into thermal imaging for automotive use 15 years ago has evolved to creating integrated thermal imaging systems that provide additional situational awareness and safety for advanced driver-assist systems (ADAS). This includes the development of robust thermal imaging datasets to enable machine vision classification in the infrared spectrum, as part of broader artificial intelligence technology stacks.
Last year, FLIR announced its Boson thermal camera technology will be part of a automaker’s level-four autonomous vehicle (AV) sensor suite through our long-time partner Veoneer. And now with the introduction of the Zoox autonomous vehicle, FLIR Systems is thrilled to see its technology continue to gain wider acceptance as a crucial component of safer mobility.
Recently, FLIR teamed up with VSI Labs to develop a proof-of-concept vehicle and testing program that looks at how fusing thermal sensor data with radar a visible sensor data via a trained neural network can improve pedestrian detection for AEB systems. By adding thermal to current sensor suites, the testing showed significant improvement in nighttime and poor visibility scenarios.
Autonomous innovation remains a strategic priority at FLIR, and as the automotive industry continues to evolve and perfect its safety systems, FLIR will play an integral role to support the future of travel.