Sensor Data Transfer Modules for Autonomous Driving Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving vehicles face inefficiencies in sensor processing due to the lack of effective sensor processing units that can handle diverse sensor types and requirements, limiting motion planning and control accuracy.
Innovation Solution
A sensor unit with a sensor interface, host interface, and multiple data transfer modules that operate in low latency, high bandwidth, and memory modes to efficiently transfer data between various sensors and the host system, accommodating different sensor types and maximizing bandwidth and processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single sensor processing unit is used to handle all sensor types, then device complexity is reduced, but processing efficiency and accuracy deteriorate due to inability to meet diverse sensor requirements
Solution Approach 1:
The sensor processing unit is segmented into multiple specialized processing channels, each optimized for specific sensor types (e.g., LIDAR, radar, camera). Each channel has dedicated data transfer modules and processing logic tailored to the characteristics of its associated sensor type, enabling efficient parallel processing while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The sensor processing unit employs a universal interface layer that can accommodate multiple sensor types through standardized connection protocols, while underlying processing channels provide type-specific optimization. This multi-functional design allows a single processing unit to handle diverse sensors efficiently without requiring separate dedicated units for each sensor type.
2Speed
If data is transferred in high bandwidth mode for all sensors, then data transfer speed is improved, but latency increases for time-sensitive sensors
Solution Approach 1:
The data transfer modules dynamically adjust their operating mode based on sensor type and data characteristics. Time-sensitive sensors (e.g., LIDAR for collision detection) automatically receive low-latency transfer treatment, while data-rich sensors (e.g., camera arrays) utilize high-bandwidth modes. This dynamic adaptation optimizes both speed and latency performance across different sensor types without manual configuration.
Solution Approach 2:
The system changes key transfer parameters such as buffer size, transfer frequency, and priority level based on sensor requirements. For critical sensors, parameters are adjusted to minimize latency (smaller buffers, higher priority), while for non-critical sensors, parameters optimize for bandwidth utilization (larger buffers, batched transfers). This parameter adaptation resolves the contradiction between speed and latency.
3Measurement precision
If multiple specialized sensor processing units are deployed for different sensor types, then processing accuracy is improved, but device complexity and resource requirements increase
Solution Approach 1:
The processing unit employs a nested architecture where general-purpose processing functions are embedded within specialized processing channels. Each specialized channel contains dedicated optimization logic for its sensor type, while also utilizing shared resources from parent processing layers. This nesting enables high processing accuracy through specialization while containing complexity through hierarchical resource sharing.
Solution Approach 2:
A centralized resource management intermediary coordinates between multiple specialized processing channels, allocating shared resources (memory, compute units, data buffers) based on real-time needs. This intermediary layer enables specialized channels to achieve high accuracy without each channel requiring complete independence, thereby reducing overall system complexity while maintaining processing precision.
4Measurement precision
If limited hardware resources are distributed across multiple sensor processing channels, then individual channel performance is improved, but overall system resource utilization decreases
Solution Approach 1:
The resource management system ensures continuous utilization of hardware resources by dynamically allocating idle resources from one processing channel to another based on real-time workload demands. When one sensor type has low data volume, its freed resources are immediately reallocated to channels with high data demands, maintaining near-continuous useful action across all resources and improving overall utilization while preserving individual channel performance.
Solution Approach 2:
The system implements resource recovery mechanisms where processing resources are temporarily released from low-priority or idle sensor channels and recovered for use by high-priority channels. This dynamic discarding and recovering of resources enables individual channels to have dedicated resources when needed (maintaining accuracy) while allowing overall system utilization to remain high through resource sharing during idle periods.
Data Source
AI summary
In one embodiment, a sensor unit to be utilized in an autonomous driving vehicle (ADV) includes a sensor interface that can be coupled to a number of sensors mounted on a number of different locations of the ADV. The sensor unit further includes a host interface that can be coupled to a host system such as a planning and control system utilized to autonomously drive the vehicle. The sensor unit further includes a number of data transfer modules corresponding to the sensors. Each of the data transfer modules can be configured to operate in one of the operating modes, dependent upon the type of the corresponding sensor. The operating modes include a low latency mode, a high bandwidth mode, and a memory mode.