Development and Evaluation of an Edge Sensor Preprocessor

Institut

Lehrstuhl für Fahrzeugtechnik (TUM-ED)

Typ

Bachelorarbeit / Semesterarbeit /

Inhalt

konstruktiv /  

Beschreibung

The aim of this practical project is to develop and evaluate a sensor pre-processing pipeline on an NVIDIA Jetson AGX edge device to offload the main compute unit of an autonomous vehicle.

MOTIVATION

Autonomous vehicles rely on a diverse suite of sensors that generate massive amounts of raw data. Typically, this data must be heavily pre-processed before it can be used for perception and planning. Common tasks include transforming LiDAR point clouds to a fixed reference frame, or converting camera color spaces and compressing video feeds. Performing all these operations centrally puts strain on the vehicle's main PC.

To address this bottleneck, this work aims to offload these tasks to an edge computing device (NVIDIA Jetson AGX). Instead of connecting sensors directly to the main PC, the Jetson AGX will serve as an intermediate processing unit. This project is highly application-focused, giving you the opportunity to build a robust, real-time hardware and software architecture that will directly improve the efficiency of the vehicle's sensor stack.

Voraussetzungen

YOUR ROLE

The work packages of this thesis contain:

• Hardware & Driver Setup: Set up the physical connections for the camera and LiDAR sensors on the NVIDIA Jetson AGX and extend and integrate the existing drivers for this new edge architecture.
• Synchronization: Integrating the new hardware into the existing PTP synchronization
• Integration & Transmission: Develop an efficient pipeline to process the data on the edge device and seamlessly transmit the pre-processed data to the main PC.
• Performance Evaluation: Conduct a thorough evaluation of the new edge node, specifically measuring system performance, latency, and transmission delays compared to the traditional centralized approach.

WHAT YOU SHOULD BRING ALONG

• Strong interest and motivation for autonomous driving and embedded systems
• Initiative and an independent, hands-on way of working
• Solid programming skills in C++ and Python
• Experience with Linux environments and basic networking concepts
• Optional but helpful: Previous experience with ROS 2, hardware integration, or NVIDIA Jetson platforms

If you are interested in joining this project, feel free to send me an application with your CV and transcript of records. I look forward to receiving your application! 

EMail: niklas.krauss<script>document.write('@');</script>

<noscript>(at)</noscript>tum.de 

 

Verwendete Technologien

Linux, Networking, Edge Computing, Autonomous Vehicles

Tags

FTM Studienarbeit, FTM Krauss, FTM AV, FTM AV Safe Operation, FTM Informatik, FTM Teleoperation

Möglicher Beginn

sofort

Kontakt

Niklas Krauß
Raum: 3507
Tel.: +49172 1736882
niklas.krausstum.de

Ausschreibung