Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An FPGA-Based Neuro-Fuzzy Sensor for Personalized Driving Assistance

Óscar Mata-Carballeira, Jon Gutiérrez-Zaballa, Inés del Campo, Victoria Martínez

TL;DR

The paper tackles personalized driving-assistance by recognizing driver DS from naturalistic SHRP2 data and mapping it to ADAS settings in real time.It combines offline DS clustering (k-means) with online, hardware-accelerated ANFIS classifiers implemented on a Xilinx Zynq PSoC, enabling rapid DS identification and THW-based ACC personalization.Key contributions include a three-cluster DS model with high identification accuracy (≈95%) and a fully FPGA-accelerated DS sensor achieving low-latency THW personalization (e.g., 0.53 µs in reported results) for real-time ADAS adaptation.Practically, the approach enables driver-specific safe margins without explicit driver intervention, improving ADAS acceptance and road safety, while maintaining low power and high-speed operation.Overall, the work demonstrates a viable, scalable pathway to deploy neuro-fuzzy DS personalization in production-grade automotive hardware.

Abstract

Advanced driving-assistance systems (ADAS) are intended to automatize driver tasks, as well as improve driving and vehicle safety. This work proposes an intelligent neuro-fuzzy sensor for driving style (DS) recognition, suitable for ADAS enhancement. The development of the driving style intelligent sensor uses naturalistic driving data from the SHRP2 study, which includes data from a CAN bus, inertial measurement unit, and front radar. The system has been successfully implemented using a field-programmable gate array (FPGA) device of the Xilinx Zynq programmable system-on-chip (PSoC). It can mimic the typical timing parameters of a group of drivers as well as tune these typical parameters to model individual DSs. The neuro-fuzzy intelligent sensor provides high-speed real-time active ADAS implementation and is able to personalize its behavior into safe margins without driver intervention. In particular, the personalization procedure of the time headway (THW) parameter for an ACC in steady car following was developed, achieving a performance of 0.53 microseconds. This performance fulfilled the requirements of cutting-edge active ADAS specifications.

An FPGA-Based Neuro-Fuzzy Sensor for Personalized Driving Assistance

TL;DR

The paper tackles personalized driving-assistance by recognizing driver DS from naturalistic SHRP2 data and mapping it to ADAS settings in real time.It combines offline DS clustering (k-means) with online, hardware-accelerated ANFIS classifiers implemented on a Xilinx Zynq PSoC, enabling rapid DS identification and THW-based ACC personalization.Key contributions include a three-cluster DS model with high identification accuracy (≈95%) and a fully FPGA-accelerated DS sensor achieving low-latency THW personalization (e.g., 0.53 µs in reported results) for real-time ADAS adaptation.Practically, the approach enables driver-specific safe margins without explicit driver intervention, improving ADAS acceptance and road safety, while maintaining low power and high-speed operation.Overall, the work demonstrates a viable, scalable pathway to deploy neuro-fuzzy DS personalization in production-grade automotive hardware.

Abstract

Advanced driving-assistance systems (ADAS) are intended to automatize driver tasks, as well as improve driving and vehicle safety. This work proposes an intelligent neuro-fuzzy sensor for driving style (DS) recognition, suitable for ADAS enhancement. The development of the driving style intelligent sensor uses naturalistic driving data from the SHRP2 study, which includes data from a CAN bus, inertial measurement unit, and front radar. The system has been successfully implemented using a field-programmable gate array (FPGA) device of the Xilinx Zynq programmable system-on-chip (PSoC). It can mimic the typical timing parameters of a group of drivers as well as tune these typical parameters to model individual DSs. The neuro-fuzzy intelligent sensor provides high-speed real-time active ADAS implementation and is able to personalize its behavior into safe margins without driver intervention. In particular, the personalization procedure of the time headway (THW) parameter for an ACC in steady car following was developed, achieving a performance of 0.53 microseconds. This performance fulfilled the requirements of cutting-edge active ADAS specifications.

Paper Structure

This paper contains 25 sections, 9 equations, 13 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Outline of Driving-Style Personalization System
  3. Driving Style Characterization in Car-Following Scenarios
  4. SHRP2-NDS Description
  5. Car-Following Situations
  6. Driving Parameters and Driving-Style Characterization
  7. Steady Car-Following Premises
  8. Car-Following Stretches in the SHRP2-NDS Trips
  9. Neuro-Fuzzy Modeling of Driving-Style Clusters
  10. Driving-Style Clustering
  11. ANFIS-Based Identification
  12. Zero-Order Takagi--Sugeno Inference System
  13. ANFIS Training
  14. ANFIS Testing and Identification Results
  15. Implementation of FPGA-Based Intelligent Sensor
  16. ...and 10 more sections

Figures (13)

  • Figure S1: Offline sequence of tasks involved in the design and development of a neuro-fuzzy sensor for advanced driving-assistance system (ADAS) personalization.
  • Figure S2: Block diagram of the field-programmable gate array (FPGA)-based intelligent sensor for online car-following ADAS.
  • Figure S3: Data acquisition systems and sensors installed in the vehicles that participated in the SHRP2-NDS. IR: infrared; SW: software.
  • Figure S4: Representative example of car-following features: (a) time-exposed time headway (TETH) and (b) time-integrated time headway (TITH).
  • Figure S5: Clusters obtained applying the k-means algorithm to the car-following segments; $\text{THW}_\text{RMS}$, $\text{TETH}$, and $\text{TITH}$ values were normalized.
  • ...and 8 more figures