Ing. Avinaash Manoharan

Deep-learning perception for automated vehicles TECHNICAL LEAD, ECOMOBILITY (May 2023 - present), EU Horizon Europe / Chips JU, ~50 partners - Created the Cumulative Consensus Score (arXiv:2509.12871) - Evaluation framework; Flask/ZeroMQ demonstrator - Supervise a working student (NVIDIA Jetson Orin) RESEARCH ASSOCIATE (Jan 2022 - Apr 2023) - Deployed CNN object detectors (PyTorch, ONNX, CUDA) on ZF ProAI and NVIDIA Jetson - ROS 2, Docker; CI/CD with Jenkins and Kubernetes

Realtime Software update (Image processing) on autonomous driving vehicles on Nvidia-ProAI - Pytorch-CUDA Framework (python) - Neural Network (Transfer Learning) - Docker - Kubernetes - Jenkins (CI-CD)

Indoor localization for Industry 4.0. UltraWideBand (UWB) localization systems are vulnerable to metals and non-Gaussian noise interference. - Robust Adaptive Kalman filter and machine learning based dynamic model creation using recursive weighted least squares, tuned for human walking scenario - Algorithm tested on python based random walk simulator Concepts: LMMSE estimator, Kalman/robust/adaptive filters, weighted least squares, particle filter, FIR filter, FFT Simulations in Matlab, Python

Project: PI Controller for a ship - create a controller for the navigation of a boat - Environment simulation using Godot Game Engine - C++ code deployment on ESP32 microcontroller - Database using InfluxDB and visualization using Grafana Indoor Positioning System - Industrial Environment Position tracking using UWB and IMU and sensor fusion Skills: Signal processing algorithms, Matlab, Octave, Embedded C, Python

Project: Analysis of Computer Vision Application - Set up AUTOSAR consortium provided demonstrator with computer vision (OpenCV) application on Raspberry Pi using Yocto build system - Combined 2 ECU images into a single ECU image using Yocto and CMake modification - LTTng tracing of Linux kernel task scheduling, displayed on INCHRON ChronView using Python/C++ library - Logging with AUTOSAR Log and Trace framework (ara::com) in C++ Skills: Linux OS, Poky (Yocto), LTTng tracing, C++

Projects: My first real time project: Internal web app development. The architecture pattern Model-View-Controller Manual software testing: Involved in creation of test cases for the software requirements. Manual test case verification and validation using HP-ALM Softwares trained on: C#.net, Java Selenium (automation testing), MySQL, C, C++ Additional languages learnt: VBscript (MS-Excel), HTML

Topic: label-free failure surfacing for object detectors - Detect where a deployed detector underperforms, without labels - Feed those cases into targeted retraining - Conducted alongside my research role at DLR Publication: Cumulative Consensus Score (arXiv:2509.12871, under review at WACV 2027)

Grade: 1.79 (1st and 2nd semester) Masters Project: Multivariable Tank Level Control - MIMO, PI controller, LQR, Kalman filter, Butterworth filter, MATLAB-Simulink VHDL/System on Chip: Zed Board (Zynq), VHDL, Vivado, Xilinx SDK; state machine, IP block and softcore FPGA design Model Based Software design: State-Flow in MATLAB-Simulink Maritime Systems: C, Atmel Studio, LT-SPICE, Matlab Industrial Systems: DC motors, induction motors, synchronous machines, frequency converters

Grade: 8.20/10 Activities: Won first prize in project presentation in several colleges, including Anna University Guindy. Bachelors Project: GSM based Automatic irrigation system using Arduino - embedded solution for irrigation system. Hardware: Arduino micro controller, SIM 900A GSM module Programming: Arduino IDE