Company attributes
Other attributes
Horizon Robotics provides hardware and software computing solutions with a focus on energy efficiency for advanced driver assistance systems (ADAS) and automated driving (AD) systems in consumer vehicles. These solutions include low-power computing hardware and open software development tools.
Journey automotive processors utilize modern neural net (NN) architectures and can be applied in driverless taxis as well as in-cabin systems. The design of Horizon Robotics' BPU is co-optimized with NN architectures such as and EfficientNet, Transformer. The BPU leverages accelerated computing technologies, including hardware and software parallelism, static random access memory (SRAM), systolic array, and concurrent compute engines.
The BPU has a computing power of 128 tera operations per second (TOPS) and is optimised for real-time AV workloads with performance at low batch size focus on latency reduction. Safety features of the BPU include error correction code (ECC) memory in SRAM, registers, buffers redundant control, and data paths.
The ADAS system's features include the following:
- Processor operating at less than 3W
- 2D/3D detection of common road objects and long tails
- More than 23 categories of semantic segmentation
- Recognition and classification of lane, vehicle, traffic light, traffic sign, and other common types of road objects.
- 1080p video input
- 30 FPS frame rate per stream
- 60 ms processing latency on the front camera
- Automatic Emergency Braking (AEB)
- Lane Keep Assist (LKA)
- Adaptive Cruise Control (ACC)
- Traffic Jam Assist (TJA)
- Automatic High Beams (AHB)
- Active Parking Assist (APA)
Demonstration of Horizon Robotics' Li L8 Pro driving assistance system
