CAN Bus Signal Analysis for Driver Abnormality Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver state detection technologies have limited bio-signal recognition rates and are unable to immediately generate statistics for fuel efficiency or average speed, making it difficult to reliably detect driver abnormalities such as drowsiness or unauthorized vehicle operation.
Innovation Solution
A method and apparatus that analyze CAN bus signals using an autoencoder to extract detection vectors and detect driver abnormalities through unsupervised learning, minimizing Mean Square Error (MSE) and using anomaly scores to determine abnormal driving conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If bio-signal recognition methods are used to detect driver state, then driver abnormalities can be detected, but the recognition rate is limited and reliability is insufficient
Solution Approach 1:
The patent replaces bio-signal recognition methods with CAN bus signal analysis. Instead of using cameras, sensors, or other devices to detect physiological signals like eye closure, facial direction, or cardiac impulses, the system analyzes driving behavior patterns from existing vehicle communication bus data, substituting a more reliable data source for the less reliable bio-signal methods.
Solution Approach 2:
The patent introduces CAN bus signals as an intermediary to detect driver abnormalities. Rather than directly observing bio-signals, the system uses driving behavior data (steering operations, accelerator/brake operations, lane changes) transmitted through the vehicle's CAN bus network as a mediator to infer driver state, providing more objective and measurable evidence of abnormal conditions.
2Reliability
If statistics for fuel efficiency or average speed are generated to detect unauthorized drivers, then vehicle operation patterns can be analyzed, but the detection is delayed until long-term statistics are accumulated
Solution Approach 1:
The patent performs preliminary analysis of driving habits by continuously learning and storing normal driving patterns during periods when the authorized driver operates the vehicle. This preliminary action creates a reference model of normal behavior that enables immediate detection of unauthorized drivers, eliminating the need to wait for long-term statistics accumulation.
Solution Approach 2:
The patent implements a dynamic detection system that adapts to different driving conditions and patterns. The system continuously updates its understanding of normal driving behavior and can immediately compare current driving patterns against this dynamic model, enabling real-time detection rather than relying on static, long-term averages.
3Measurement precision
If additional bio-signal recognition equipment is installed to improve detection accuracy, then driver state can be monitored, but device complexity and cost increase
Solution Approach 1:
The patent makes the existing CAN bus system serve multiple functions. The same communication network that handles routine vehicle operations (engine control, transmission, braking) is also utilized for driver abnormality detection, eliminating the need for separate dedicated detection hardware and reducing overall system complexity.
Solution Approach 2:
The patent enables the vehicle's existing communication infrastructure to serve its own monitoring needs. The CAN bus network, which already collects data about vehicle operations, now also provides data for driver state analysis without requiring additional sensors or communication channels, making the system self-sufficient.
Data Source
AI summary
Provided is a method for detecting a driver's abnormalities based on a CAN (Controller Area Network) bus network communicating with an ECU (Electronic Control Unit) of a vehicle. The method may include: acquiring a CAN bus signal related to an operation of the vehicle from the CAN bus network; extracting a detection vector from the CAN bus signal using an auto encoder; and detecting a driver's abnormality based on the detection vector.