Sensor cleaning systems from IAV
for autonomous vehicles
Autonomous and highly automated vehicles – such as UGVs and other special-purpose platforms - rely robust, precise sensor systems. Cameras, IR cameras, LiDAR, radar and further sensor systems form the foundation for environment perception and functional safety.
A key prerequisite is permanently unobstructed visibility for the sensors – even in rain, dust, mud, dirt, soot, smudges, ice or snow.
Without reliable cleaning of cameras and other sensors, image quality, sensor performance and system availability of automated driving or other sensor based functions deteriorate significantly due to soiling and other perturbations such as waterdrops and raindrops.
IAV’s sensor cleaning system addresses this challenge with an integrated approach combining AI-based soiling detection and targeted, automated cleaning of exposed sensors. This sustainably supports availability, functionality and operational safety of vehicles both on-road and off-road, even under heavy soil loads.
Modular sensor cleaning system for cameras and other sensors
- AI-based modules for real-time soiling detection
- Components for targeted camera and sensor cleaning (nozzles, valves, pumps, media supply)
- System architecture and integration concepts for different vehicle platforms
A scalable system platform can cover a wide range of sensors – from cameras and IR cameras to LiDAR and other exposed sensor systems. The modular design enables flexible adaptation to platform, sensor layout, use case and production volume.
The system detects functionally relevant soiling and other perturbations and initiates appropriate cleaning actions on demand. Harmonized components for detection, actuation and media supply support high availability of cameras and sensors under varying environmental conditions. Individual modules can be scaled, combined or customized depending on the application.
Engineering & integration – sensor cleaning in the vehicle
The use of camera and sensor cleaning systems requires close coordination with vehicle architecture, sensor layout and the intended operating scenario. IAV supplements the system toolkit with engineering and integration services, including:
- Derivation and definition of system requirements
- Design of the cleaning architecture and media supply
- 3D CFD simulation of soiling and of the cleaning system itself
- Alignment with camera and sensor positions as well as packaging
- Integration into existing vehicle and platform concepts
This results in application-optimized camera and sensor cleaning – from initial functional samples through to series-ready system concepts.
AI-based soiling detection for cameras and LiDAR
The basis for on-demand camera and sensor cleaning is reliable detection of functionally relevant soiling and signal perturbations. IAV uses AI-based methods to analyze sensor data, for example from camera or LiDAR systems, in order to identify impairments to sensor perception at an early stage.
The algorithms detect various types of image and signal artifacts caused by rain, raindrops, water, waterdrops, spray, dust, dirt, mud, snow, ice and other environmental influences (including smudges on lenses), and they support situation-dependent cleaning strategies – for example by adapting cleaning intensity, duration or the amount of cleaning media.
In multisensor architectures, information from several sensor sources can be fused. This cross-sensor evaluation increases the robustness and stability of soiling detection and perturbation assessment – a key foundation for automated and autonomous driving functions.
From proof of concept to series production
The modular sensor cleaning system is designed for different maturity levels and volumes:
- Proof of concept and advanced development to evaluate cleaning strategies, system architectures and integration concepts
- Prototypes and small series to test camera and sensor cleaning in the field
- Scalable, near-series system concepts for high-volume production and platform strategies
Insights from early project phases are gradually incorporated into production-ready solutions and enable a consistent development path.
Mobile demonstrator – sensor cleaning in real-world operation
IAV provides a mobile demonstrator to showcase the interaction between AI-based soiling detection and sensor cleaning under realistic conditions.
The demonstrator:
- visualizes the impact of different soiling situations on real cameras
- shows the reaction of the cleaning actuators in real time
- serves as a platform for discussing and evaluating system concepts and integration issues
Together with customers’ specialist teams, requirements, architecture variants and next steps along the individual project path can be derived – from initial evaluation through to near-series applications.