Reliable and cost-effective 360° vision creation, enabled by various vision and non-vision sensors, is at the heart of development and commercialization of autonomous vehicles. With the strong push for ADAS and autonomous vehicles, sensor market is projected to grow from $3 billion in 2015 to $35 billion by 2030 [1]. One of the key autonomous functionality-enabling sensor is LIDAR (Light Detection and Ranging) that was first used in the concept vehicles competing in DARPA (Defense Advanced Research Project Agency) Grand Challenge for autonomous vehicles [2]. LIDARs rely on time of flight measurement for emitted light from laser diodes to be detected by receivers after reflecting back from objects in its path, and thus provide highly accurate information not only about object distance but also offer 3D information on object width and height.

One of the early pioneers of LIDAR technology is Velodyne that introduced mechanical rotating LIDARs to offer 360° vision. The mechanical rotating LIDARs are the most distinguishing visible hardware on the vehicle bodies for the autonomous test fleet on the road today. Although such LIDARs provide 360° view with very high resolution and long range (up to 200 m), high cost is a key deterrent for these LIDARs to be part of commercially-viable autonomous vehicles. Although in recent years, some companies, most notably Waymo [3], have made significant advances to significantly reduce cost of these mechanical rotating LIDARs. Additionally, large size and potential reliability issues due to mechanical parts, pose additional challenges for vehicle integration. Due to these issues, industry is moving towards solid-state LIDARs which promise to be in the $100-$250 LIDAR price point at large scale volume. Solid-state LIDARs don’t have costly electric motors as used in mechanical rotating LIDARs thus offer more cost-effectiveness but suffer from limited field of view (FoV) and lower range/resolution. Multiple solid-state LIDARS will be needed to be integrated in autonomous vehicle. A large number of companies developing solid-state LIDARs are pursuing innovative emitter/detector technologies to improve range and resolution as well as component and functionality consolidation to achieve desired miniaturization.

Design goals for LIDARs (solid-state or mechanical rotating ones) are largely centered on size and cost reduction without sacrificing (or increasing) range and resolution. Increase in optical power benefits LIDAR range whereas integration of multiple light emitters and detectors on single monolithic chip sets and higher signal processing improves resolution. These design factors combined with desired small form factors may cause significant heat buildup that may be detrimental to performance and durability of LIDARs. This may deter LIDAR size (and cost) reduction efforts. Additionally, LIDARs when integrated in vehicles must function reliably in an automotive environment and in all-weather conditions. Vehicle mounting location may further present additional thermal challenges that sensor designers will need to account for while designing sensors. The fact that different auto makers are looking to integrate LIDARs in different parts of vehicle body, for instance side bumpers, front grill, headlights or taillights etc, further enhances design complexity for sensor vendors.

This white paper is aimed at highlighting thermal-driven design challenges for LIDARs as well as impact of vehicle integration strategy on reliability in real-world operation. Underlying thermal issues resemble that in any consumer electronics, however, for LIDARs and other autonomous vehicle sensors unique challenges arise because of much harsher automotive environment and the associated reliability and safety issues with these sensors. We’ll showcase how Mentor’s EDA-centric electronic cooling simulation software can be exploited by sensor vendors and Auto OEMs alike to account for these challenges. For sake of consciousness, we’ll take a generic solid-state LIDAR example but the challenges highlighted here apply equally to mechanical rotating LIDARs as well. Key takeaway message for readers is that thermally-conscious designs for LIDAR signal processing electronics as well as for their enclosures, while taking into account their vehicle integration locations, is critical to ensure desired size, cost, performance and life goals are met.