Sensor Plugin Architecture for Reliable Multi-Sensor Grid Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles rely on grid maps that integrate sensor information from various sensors with different characteristics, necessitating a method to normalize and standardize this information for effective vehicle control.
Innovation Solution
A sensor plugin architecture is employed to process sensor information from multiple sensors, including radar, lidar, and cameras, by determining modality, performing plausibility checks, and integrating it into a multi-layer grid map based on sensor-specific processing and calibration, ensuring accurate vehicle control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If sensor information from multiple sensors with different characteristics is integrated directly into the grid map, then the quantity of sensor data is increased, but the reliability and accuracy of the grid map deteriorates due to lack of normalization
Solution Approach 1:
The sensor plugin architecture segments the sensor processing pipeline into distinct modular components: sensor information reception, modality determination, plugin instance generation, sensor-specific processing, plausibility checking, and grid map integration. Each sensor type (radar, lidar, camera) has its own dedicated processing chain, allowing independent normalization and validation before integration into the unified grid map structure.
Solution Approach 2:
The system transforms sensor information by applying sensor-specific parameters and processing rules based on modality. Each sensor type undergoes parameter-adjusted processing tailored to its characteristics (e.g., radar velocity data, lidar spatial data, camera image data), converting diverse sensor outputs into a standardized format suitable for grid map integration while preserving the unique informational value of each sensor type.
2Reliability
If sensor information is processed and normalized before integration, then the reliability of grid map is improved, but the complexity of the processing system increases
Solution Approach 1:
The sensor plugin architecture implements a universal processing framework that handles multiple sensor types through a common interface and standardized workflow. The multi-instanceable sensor plugin system provides a unified mechanism for modality determination, processing parameter selection, and grid map integration that works across all sensor types, reducing overall system complexity despite the diverse processing requirements of individual sensors.
Solution Approach 2:
The sensor plugin architecture acts as an intermediary layer between raw sensor information and the grid map integration system. This intermediate processing stage performs normalization, validation, and format conversion, mediating between the diverse characteristics of multiple sensors and the standardized requirements of the grid map structure, thereby simplifying the overall integration process.
3Measurement precision
If plausibility checks and sensor-specific processing are performed, then the accuracy of sensor information integration is improved, but the processing time increases
Solution Approach 1:
The system performs preliminary processing actions including modality determination, appropriate plugin instance generation, and sensor-specific preprocessing before the actual grid map integration. Plausibility checks are conducted in advance on sensor information characteristics, allowing the system to validate data quality and select appropriate processing parameters beforehand, thereby streamlining the subsequent integration process and reducing overall processing time.
Data Source
AI summary
Systems and methods for controlling an autonomous vehicle. One system includes an electronic processor and a memory configured to store a multi-layer grid and a multi-instanceable sensor plugin. The multi-layer grid includes information associated with positions in the environment. A plurality of sensors are configured to output sensor information. The electronic processor is configured to, for each of the plurality of sensors, determine a modality of the sensor and generate an instance of the sensor plugin based on the modality. The electronic processor receives, processes, and determines a plausibility of the sensor information. In response to determining that the sensor information is plausible, the electronic processor integrates the sensor information to the multi-layer grid. The electronic processor controls vehicle movement based on the multi-layer grid.