Kushal Narasimha

* Type approval of automotive active safety systems such as Driver Control Assistance Systems (UN R171), Automated Lane Keeping Systems (UN R157) and Automated Driving Systems. * Assessment of Functional Safety (ISO 26262) and Safety of Intended Functionality (ISO 21448) of the system.

* Designed and implemented Simultaneous Localization and Mapping (SLAM) algorithm and utilized CarMaker for virtual test driving of autonomous cars, taking into account Formula Student Germany competition tracks. * State Estimation of velocity and position of autonomous car using Extended Kalman Filter (EKF) by sensor fusion of Camera and IMU.

Developed a transformer-based 3D object detection and tracking model using LiDAR sensor data, which can be utilized to generate realistic scenarios in Advanced Driver Assistance Systems (ADAS) / Automated Driving (AD) simulations.

* Implemented a function to save the state of Deep Reinforcement Learning agent (PPO), which has been trained and tested on huge Time Series Data from Electrical Power Transmission and Distribution Grids. * Required Internal state variables, Parameters and Hyperparameters of the Deep RL agent were analyzed and save/load function is built in order to save the agent state at any point in time.

* First the KITTI data is converted to rosbag using kitti2bag converter then lidar point clouds are mapped into semantic segmented image. * For each pixel in every image frame, the semantic class of car and pedestrian are clustered. Later the BBox are drawn around these cluster classes. Detection, image and velodyne points are visualized in ROS RViz.

* Implemented a camera and radar based forward collision warning system which calculates the ‘time to collision’ (ttc) to the closest vehicle in the same lane as the ego vehicle and gives a warning if the ttc is below a threshold using BASELABS Create Embedded software library. * The GetTimeToCollision function is used for the implementation. The ttc is then displayed below the camera image and the background turns red if the ttc is smaller than 20 seconds.

Specialization: Computer Vision | Multisensorial System | Smart Sensor Systems | Real Time Operating Systems | Digital Signal Processing | Design of Software for Embedded Systems | Hardware/Software Codesign | Digital Components and Architecture | Digital and Mixed Signal testing | Design of Digital Systems. Current grade: 1.7/5 (German Grading System)

Specialization: Engineering mathematics | Engineering physics | Control systems and modern control theory | Programming in c and data structures | Signals and systems | Advanced power electronics | Measurements and instrumentation. Grade: First Class with Distinction (Indian Grading System), 1.7 / 5 (equivalent German Grading System)