Electronics Weekly
According to research consultancy Markets and Markets, the ADAS (advanced driver assistance systems) sector is expected to grow from 334 million units in 2024 to 655 million units in 2030.
As more radar sensors are added, sensor data needs to be managed effectively to ensure safe, real-time decision-making. (Unlike cameras or lidar, radar works effectively in all weather conditions.)
The growth in complexity of ADAS and autonomous driving functionality has seen today’s vehicles having up to five radar streams on the front, rear and side of the car, providing a 360o view.
There are multiple benefits to having more radar sensors. They detect objects from multiple angles and using more sensors provides redundancy and reduces blind spots. This is crucial for features such as lane departure warnings and adaptive cruise control.
Combining data from multiple radar sensors vehicles can create a more detailed understanding of the environment. This enhances accuracy when tracking moving objects, for example, pedestrians, cyclists and other vehicles.
Placing radar sensors at the corners of vehicles can provide cross-traffic alerts, a feature promoted by the New Car Assessment Program organisation and other safety bodies.
Managing multiple radar sensors
Each radar sensor generates large amounts of data by continuously scanning the environment and this data must be processed to ensure timely decisions.
The amount of data being transmitted to the central processor can easily overwhelm the vehicle and create the risk of bottlenecks where critical information may be delayed or missed.
The data needs to be combined with input from other sensors, including cameras and lidar, for a vehicle to make sense of its environment. This fusion requires significant computational power and improper handling can lead to latency or inaccuracies in decision-making. Processing large volumes of radar data can be power-intensive, which is especially problematic for electric vehicles.
As radar sensors become more common, the risk of interference between vehicles increases, potentially resulting in incomplete or inaccurate data.
Data challenge
Vehicle manufacturers are adopting new approaches to vehicle architecture to overcome these challenges. They include centralising and streamlining data processing using technologies such as radar bridges to provide infrastructure to manage the growing number of sensors without compromising performance or safety.
A radar bridge is a dedicated processor designed specifically to handle the output signals from multiple radar sensors. The sensors have overlapping fields of view, creating a detailed perception of the environment. This improves the vehicle’s ability to detect and track objects from multiple angles. The radar bridge is dedicated solely to processing radar data, so delay is minimal, ensuring the vehicle can react to its environment in real time.
It also helps with interference management; it can analyse data from multiple sensors simultaneously and reallocate frequencies or timeslots to avoid signal overlap. This reduces the risk of misinterpretation due to interference.
Decentralised processing
Vehicle manufacturers are adopting zonal ECUs (electronic control units) as part of a more efficient vehicle architecture. These are strategically placed throughout the vehicle to process data locally before transmitting the most relevant information to the central ECU.
Zonal ECUs reduce the amount of data that needs to be transmitted to the central processor, alleviating potential bottlenecks and improving overall system efficiency. It also alleviates the complexity of the vehicle’s wiring harness, which currently spans several kilometres in length and adds significant weight and cost to vehicle production.
Zonal ECUs also make it easier for manufacturers to enable over the air software updates in a software-defined vehicle.
