Nisarg Upadhyay Email & Phone Number

Hyattsville, Md, Us

College Park, Maryland, United States

• Tested and developed the ROS 2 Nav2 package for a warehouse robot handling autonomous navigation and air quality monitoring. Utilized the SLAM Toolbox for real-time mapping and the Robot Localization package for improved positional accuracy. Simulated the warehouse in Isaac Sim to improve testing efficiency.• Improved localization accuracy in dynamic warehouse environments by implementing a sensor fusion algorithm that integrates wheel encoder data with IMU readings using an Extended Kalman Filter (EKF). Collaborated in assembling and integrating all sensors on the robot to ensure optimal performance.• Integrated depth-enhanced object detection using the YOLOv8 model and depth maps with the Nav2 planner on the Nvidia Jetson AGX Orin, achieving a 30% improvement in navigation efficiency and a 75% reduction in computational overhead compared to RTAB-Map in real-time object detection and path planning.

College Park, Maryland, United States

• Researched the Event Camera for capturing scenes with high temporal resolution, leveraging its motion tracking and light intensity detection capabilities toachieve accurate real-time surface reconstruction and enhanced 3D normal flow analysis in rapidly changing environments compared to traditional cameras.• Developed algorithms using the EVIMO2v2 dataset to obtain the Normal and Normal Flow from the Point Cloud to fit the surfaces.• Processing time was decreased from the previous 10 to 15 seconds required using the PointNet Architecture to about 1 to 1.2 seconds each time slice by using the Pyramid technique on the Point Cloud. This improvement allowed for real-time operation and significantly increased the algorithm's performance.

United States

• Working on heavy vehicles such as Buses and Trucks for repair and maintenance.• Servicing Turbos, Brakes, Alternators, and Fluid systems along with Cummins, Ford, GM engines and, Voith, Allison transmissions.

Vadodara, Gujarat, India

• Conducted data preprocessing, cleansing, and classification on extensive datasets from diverse US retailer cooperations, leveraging a spectrum of Machine Learning Algorithms.• Applied algorithm selection methodologies, considering dataset attributes and target outcomes, to discern the optimal models for distinct data types.• Predicated customer interaction towards certain brands based on seasonal data with 93% accuracy using DBSCAN and hierarchical clustering.

Anand, Gujarat, India

• Worked as an instructor, helping students understand Image Processing and Computer Vision theory while assisting with programming assignments.

Master'S Degree, Robotics Technology/Technician

Bachelor Of Technology, Mechatronics, Robotics, And Automation Engineering