Autonomous Vehicle Fault Detection System

AUTONOMOUS VEHICLE
FAULT DETECTION

Cloud-based diagnostic system for Local Motors' Olli autonomous shuttle, revolutionizing vehicle maintenance through intelligent CAN bus data analysis

97% REDUCTION IN DIAGNOSTIC TRIAGE TIME

Local Motors Olli

The Olli is a 3D-printed, electric autonomous shuttle designed for low-speed urban environments. Our fault detection system monitored its comprehensive CAN bus network to ensure optimal performance and safety.

This 12-passenger vehicle operates on campuses, hospitals, and city centers, making real-time diagnostics critical for passenger safety and operational efficiency.

Vehicle Type Autonomous Electric Shuttle
Capacity 12 Passengers
Max Speed 25 mph
Range 100 miles per charge
Construction 80% 3D-printed components

SYSTEM ARCHITECTURE

01

CAN Bus Integration

Real-time monitoring of all vehicle diagnostic data through Controller Area Network protocol, capturing fault codes, sensor readings, and system status across all vehicle subsystems.

  • Python
  • CAN Protocol
  • Real-time Processing
02

Cloud Infrastructure

Scalable AWS deployment with automated data pipeline, fault classification algorithms, and intelligent routing to appropriate engineering teams based on fault severity and type.

  • AWS EC2
  • Lambda Functions
  • RDS Database
03

Smart Assignment

Machine learning algorithms analyze fault patterns and automatically assign issues to specialized teams, eliminating manual triage and reducing response time by 97%.

  • Classification Algorithms
  • Pattern Recognition
  • Automated Routing
04

Role-Based Dashboards

Customized interfaces for engineers, technicians, and development teams, providing relevant fault data, historical trends, and actionable insights for each role.

  • React.js
  • Data Visualization
  • Role-based Access

DATA FLOW

1
CAN Bus Data Capture
Vehicle diagnostic messages collected in real-time
2
Cloud Processing
AWS-based analysis and fault classification
3
Intelligent Routing
Automated assignment to specialized teams
4
Team Notifications
Real-time alerts and dashboard updates

PROJECT IMPACT

97%
Triage Time Reduction
100%
CAN Fault Coverage
24/7
Real-time Monitoring
3
Team Dashboards

Technical Implementation

The system was built using Python for CAN bus communication and data processing, deployed on AWS infrastructure for scalability and reliability. Machine learning algorithms were trained on historical fault data to enable intelligent classification and routing of new diagnostic issues.

Custom dashboards were developed using modern web technologies, providing role-specific views for different team members. The system successfully detected all CAN-listed vehicle faults and reduced manual diagnostic triage time from hours to minutes.