Hridik Hingorani

Robotics, controls, embedded systems. The interactive version of this site needs JavaScript; the projects are listed below.

Completed work first. In-progress and planned items are labeled.

Real-Time Task-Space Impedance Control, CRS Arm · 1 kHz task-space impedance controller in C on a TMS320F28335 DSP, joint torque commands to a 3-joint arm, 5 N·m saturation
UR3 Vision-Guided Contour Drawing · from-scratch closed-form 6-DOF inverse kinematics, product-of-exponentials forward kinematics, Canny and contour extraction, 142-stroke drawing in 7 minutes
RISC-V Preemptive Multitasking OS Kernel · roughly 6,000 lines of C, preemptive scheduler, Sv39 three-level paging, VirtIO, FAT, pipes, shell · code available on request
Autonomous Vehicle Perception and Control · in progress · perception, LiDAR sensing, reinforcement learning, and vehicle control for a self-driving stack on a Polaris GEM platform
Humanoid RL Policies in Simulation · in progress · RL policies for locomotion, balance, and manipulation; Isaac Gym and MuJoCo on CUDA, viser for visualization
Flex-Sensor Teleoperated Hand · in progress · 5-channel flex-sensor board routed in KiCad, ESP-NOW leader-follower link, PCA9685 servo output
Animatronic Grogu · planned · 22-DOF bipedal animatronic with an LLM voice pipeline (ECE 445 capstone)
Jupiter AI · fully local voice assistant on Apple Silicon; Whisper STT, Qwen3 on MLX, Kokoro TTS, real barge-in
Inverted Pendulum, Pole Placement and Observer · pole-placement state feedback with a Luenberger observer as two decoupled 2-state blocks, gain-scheduled across upright and hanging linearizations, friction parameters from logged step data, 1.1 s settling time
Cart-Pole Control Simulator · Rust, scratch LQR via discrete Riccati iteration, 6 gain presets benchmarked against tuned PID, RK4 at 1 kHz
Motion Planning Visualizer: A* vs RRT* · from-scratch A* and RRT*, A* 6 to 11% shorter paths under 20 ms

Resume (PDF) · Full CV (PDF) · hridikh2@illinois.edu