At ford, I developed, tested and deployed a custom ADAS that enhanced safety.
Problem Statement
Accidents involving Material Handling Vehicles cause significant damage to life and property, and yet they continue to remain an indispensable part of the logistics chain.
This project aimed to solve the problem by developing a custom Driver Assistance System that could retro-fit onto any MHV with an aggressive price target of 300$/unit.
My Contribution
Detailed driver behavior research analysis via interviews.
Developed a novel sensor-fusion logic, integrating RADARs and SONARs.
Conducted over 40hrs of HIL testing.
Created detailed wiki to accelerate future development cycle.
Impact
Deployed MVP in just 11 weeks.
Enhanced driver awareness, ergonomics & safety.
Concept was adopted for further development and larger pilot.
The Process
Problem Analysis
Conducted on-site interviews to understand driver behavior and root cause of errors. Co-designed feedback mechanisms to enable seamless integration into their existing workflow.
User Analysis | Co-Designing
Sensor - Fusion
Conducted in-depth research on various sensors and processors, evaluating their computational requirements.
Developed a novel sensor-fusion approach integrating RADAR, SONAR, and accelerometers for advanced vehicle state estimation, enabling optimized sensing patterns for enhanced performance
FMCW RADARs | SONARs | Component Research
Prototyping
Prototyped the entire system in a shoe-box on a mobile hand-cart, testing and collecting indoor data for optimization and calibration.
Developed C++ based firmware to enable intra-processor communication using USB as the central communication protocol and UART/IIC/SPI on the edge.
C++ | Python | Embedded Systems | Sensor Calibration
Feedback Mechanism
User research highlighted a need for a new feedback mechanism which the the drivers "can view off of the corner of their eye".
Taking inspiration from blind-spot monitors and studying human peripheral vision processing capabilities, I developed a custom LED matrix display.
Information Architecture | Embedded Firmware | UI Design
Final MVP
Mounting
Designed a magnet-based enclosure enabling versatile and customizable mounting.
Smart Monitoring
Used a multi-colored status bulb to monitor key sensor and processor status.
