Sensor Data Prioritization for Automotive CAN Network Jitter Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing sensor systems in autonomous vehicles face delays and resource inefficiencies due to unorganized data transmission, leading to 'jitter' in the CAN network, which disrupts the timing and accuracy of data representation, impacting reaction times in autonomous driving scenarios.

Innovation Solution

A method that uses a common reference frame for all sensors to convert and prioritize data transmission based on object proximity and nature, employing a dynamic CAN-ID assignment to optimize data organization and reduce jitter, ensuring efficient communication through the CAN network.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data from multiple sensors are transmitted without organized priority on the CAN network, then all sensor data can be communicated, but jitter occurs in the network which disrupts timing and increases processing complexity

Engineering Contradiction:
Improvedata transmission reliabilityVSAvoiddata processing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by organizing and prioritizing sensor data in buffer memories before transmission on the CAN network. Data are sorted according to object proximity and nature characteristics prior to communication, which eliminates jitter and simplifies subsequent processing at the control module without requiring complex real-time prioritization algorithms.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If a likelihood algorithm is used to verify object detection across multiple sensors, then object identification accuracy improves, but CPU time and RAM memory resources are significantly consumed

Engineering Contradiction:
Improveobject identification accuracyVSAvoidCPU and RAM resource consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent reduces the computational burden of the likelihood algorithm by preliminarily organizing sensor data according to object characteristics before transmission. This pre-organization filters and structures the input data, allowing the likelihood algorithm to operate on already-sorted information, thereby reducing CPU time and RAM memory requirements while maintaining identification accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent applies local quality by assigning different priority levels to different types of objects based on their characteristics. Critical objects such as pedestrians or cyclists receive higher priority in data transmission and processing, allowing the system to concentrate computational resources on the most important detections rather than uniformly processing all objects with equal resource allocation.

Inventive Principle:
Principle #3Local quality

3Adaptability or versatility

If sensors use their own local reference frames to locate objects, then each sensor can independently detect objects, but data integration and verification across sensors becomes more difficult

Engineering Contradiction:
Improvesensor independenceVSAvoiddata integration complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a common reference frame as an intermediary for integrating data from multiple sensors that operate in their own local reference frames. By converting sensor data into a unified reference system based on object characteristics, the patent enables straightforward data integration and verification across sensors while preserving the independence and adaptability of individual sensor detection capabilities.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3613196B1Processing information acquired by a plurality of sensors of an automotive vehicle
Publication Date: 2023.10.25 STELLANTIS AUTO SAS
  • EP3613196B1 patent drawingFigure 1~2
  • EP3613196B1 patent drawingFigure 3~4
  • EP3613196B1 patent drawingFigure 5~6

AI summary

Method for processing information arising from a plurality of sensors (Cx) on board an automotive vehicle, said sensors (Cx) each comprising means (M1) for acquiring data relating to one or more objects (OA) present in the environment of the vehicle and located in a determined frame of reference specific to each sensor (Cx) and at various instants, means (M2) for processing said data and means (M3) for communicating the data comprising several buffer memories (BF1-BFn) in which the data are stored before their transmission to a control module via a communication network (CAN) of the vehicle, said method being characterized in that it consists in using one and the same common frame of reference for the object or objects (OA) located by each sensor (Cx) and in organizing the buffer memories (BF1- BFn) of each sensor (Cx) as a function of one or more characteristics or criteria of the object or objects (OA) located by each sensor (Cx) in the common frame of reference; said characteristics or criteria being used to define an order of priority of transmission of the outgoing data exiting each sensor (Cx) on the network (CAN).