Aneesh Reddy Vallapureddy

Designed and implemented a U-Net-based semantic segmentation model using Python and TensorFlow to classify road objects in self-driving car environments.
Processed real-time video data to detect and segment key features such as lanes, vehicles, and pedestrians, improving object recognition and navigation accuracy.
Optimized model performance through hyperparameter tuning, data augmentation, and computational efficiency enhancements, enabling real-time operation.
Collaborated with a team to integrate the segmentation model into a simulated autonomous driving framework.

Python TensorFLow U-Net PyTorch OpenCV ROS Ubuntu

Conceptualized a smart warehouse automation system integrating motion-controlled conveyors and vision-based object recognition to optimize storage and retrieval processes.
Explored the use of OpenCV for barcode detection, item classification, and defect identification to automate sorting and quality assurance workflows.
Designed a prototype Human-Machine Interface (HMI) using Ignition SCADA, incorporating inventory visualization, real-time status updates, and automated error alerts to enhance system monitoring and control.
Modeled conveyor system dynamics and load management strategies using MATLAB Simulink, proposing motion control mechanisms for efficient item transportation.

Ignition Matlab/Simulink OpenCV AnyLogic

Developed and implemented sensor-driven rotor balancing procedures, utilizing IoT-enabled sensors to monitor real-time vibration and rotational dynamics, enhancing machine stability and operational efficiency.
Integrated real-time data acquisition and analysis pipelines, employing advanced signal processing techniques to detect imbalances and predict potential failures, aligning with Industry 4.0 standards for smart manufacturing.
Collaborated with senior engineers to interpret sensor data and derive actionable insights, identifying critical areas for automation and recommending optimization strategies to improve machinery performance.
Earned commendation for innovative contributions, demonstrating the practical application of sensor technologies and data analytics to enhance industrial automation and reduce downtime.

IoT SciPy NumPy Industry 4.0

University at Buffalo

MS Engineering Science (Robotics) - August 2023 - December 2024

Mahindra University

B.Tech Mechanical Engineering- August 2019 - May 2023

Implemented pure pursuit, RRT, A*, and gap-following algorithms for path planning, obstacle detection, and lateral vehicle control.
Utilized stereo visual odometry, sensor fusion (LiDAR and IMU), and RANSAC for accurate vehicle localization, obstacle detection, and environment mapping.
Developed and visualized local occupancy grids for real-time navigation, optimizing path planning in unmapped and dynamic environments.
Applied probabilistic localization techniques, including Bayesian Estimation and Particle Filters, ensuring robustness under noisy and uncertain conditions.
Simulated and validated navigation strategies in the F1Tenth Simulator, achieving efficient trajectory tracking and obstacle avoidance in complex scenarios.

ROS python OpenCV RRT F1Tenth RViz path planning

Designed and implemented advanced control algorithms, including inverse kinematics and PID control, for precise joint-level manipulation of the UR-3 robotic arm in pick-and-place operations.
Integrated real-time sensor feedback for dynamic motion adjustments, leveraging data from vision systems and tactile sensors to enhance task accuracy and system stability.
Developed motion planning strategies using ROS-based MoveIt, including collision-aware trajectory generation and workspace optimization for efficient robot operations.
Utilized OpenCV for vision-based object detection and alignment, enabling accurate object localization and orientation correction within the robot's workspace.
Simulated and validated the control algorithms in URSim and MoveIt before deployment, ensuring seamless operation in real-world scenarios.

Matlab MoveIT URSim ROS python openCV

Developed a real-time facial emotion recognition system using a Convolutional Neural Network (CNN) trained on the FER-2013 dataset, achieving emotion classification across seven emotional states: happy, sad, neutral, fear, angry, surprise, and disgust.
Preprocessed data with techniques such as grayscale conversion, image normalization, and one-hot encoding, ensuring consistent input for model training and improved convergence.
Optimized the CNN model by tuning hyperparameters like learning rate and batch size, leveraging advanced layers like BatchNormalization, MaxPooling, and Dropout for regularization.
Enhanced feature extraction through data augmentation and Canny edge detection, increasing model generalization and robustness to varying conditions such as lighting and facial orientations.
Deployed the model for real-time testing with live webcam feeds, enabling dynamic emotion detection to enhance user interaction in social robotics applications.

CNN Keras/TensorFlow python OpenCV

Developed a robust lane detection pipeline leveraging Digital Signal Processing (DSP) techniques and advanced computer vision methods to enhance lane visibility and identification under diverse driving conditions.
Implemented preprocessing techniques, including Gaussian filtering, Canny edge detection, and Contrast Limited Adaptive Histogram Equalization (CLAHE), to improve noise reduction, contrast enhancement, and feature extraction for real-time application
Integrated adaptive thresholding and perspective transformation, enabling dynamic adjustments to fluctuating illumination and providing a bird's-eye view of the roadway for enhanced lane detection.
Utilized sliding window algorithms combined with polynomial fitting for accurate tracking of straight and curved lanes, improving system performance on challenging road geometries such as broken or faded markings.
Achieved reliable real-time lane visualization by overlaying detected lane regions, contributing to autonomous driving and Advanced Driver Assistance Systems (ADAS) for safer navigation and adaptive vehicle control.

Engineered a framework to retrofit the Intelitek Promill 8000 CNC machine with smart IoT sensors, enabling real-time monitoring of parameters such as feed rate, tool position (X, Y, Z axes), vibration, coolant levels, and cutting forces.
Integrated diverse sensors, including ultrasonic, infrared, accelerometer, gyroscope, and temperature sensors, for robust data acquisition and machine status analysis.
Developed a data acquisition pipeline using Raspberry Pi and NodeMCU to collect and process sensor data, transmitting it to a cloud-based Google Firebase database for remote access and storage.
Implemented real-time dashboards and digital twins for data visualization, enabling shop-floor managers to monitor machine performance and optimize machining operations remotely.
Achieved significant transformation by converting the Industry 3.0 CNC machine to an Industry 4.0-ready system, facilitating predictive maintenance, performance tracking, and efficient resource utilization.

IoT Database python RaspberryPi

Robotics Engineer