Ankit Aggarwal

Carnegie Mellon University

Pittsburgh, PA

I architect autonomous systems designed to thrive where traditional engineering fails: unpredictable, unstructured, and extreme environments. From my time leading Mars Rover Manipal to delivering the Lunar ROADSTER at Carnegie Mellon, my path has been defined by a from-scratch philosophy. I bridge the gap between high-level intelligence and rugged physical execution. I enjoy the challenge of working across the entire robotics stack ranging from deriving the mathematical foundations for a deep neural network to prototyping custom cycloidal gearboxes to ensure hardware resilience.

I thrive in the 0-to-1 phase of robotics, taking a complex, multidisciplinary problem and building a robust, field-ready solution. I am currently seeking Founding Robotics Engineer or early-stage technical roles where I can own the end-to-end development of complex systems and work across the robotics stack.

Technical Skills

I am a full-stack engineer capable of owning a project from the mathematical derivation of a control law to the physical fabrication of the chassis.

The Body: Mechatronics and Hardware

Precision Mechanical Design: SolidWorks Professional (CSWP) certified. Specialized in custom cycloidal gearbox design, tendon-driven systems, and mechanical logic elements for reliable control.
Embedded Development: HAL/LL/Register-level programming for STM32 and TI Launchpad. Skilled in real-time firmware (C/C++), low-latency communication (SPI, I2C, CAN, UART), and industrial PLC interfacing via Python RPCs.
Power and Fabrication: Integration of BLDC actuators, motor drivers, and battery management systems. Experienced in 3D printing for mechanical assemblies and hardware-in-the-loop (HIL) testing.
Circuit and PCB Design: Expertise in schematic capture and PCB layout for sensor interfacing and power distribution. Proficient in component selection, mixed-signal systems, and rapid prototyping of custom breakout boards.

The Brains: Autonomy, Controls, and Planning

Motion Planning: Proficient in search-based (A*, Dijkstra) and sampling-based (RRT*, PRM) planners. Experienced in lattice-based planning and navigation in unstructured off-road environments.
Control Theory: Expertise in PID, Model Predictive Control (MPC), and Inverse Kinematics (IK). Skilled in trajectory optimization and state estimation for high-degree-of-freedom manipulators.
Robotics Foundations: Deep understanding of rigid body dynamics, spatial transformations, and physics-based simulation using Isaac Sim and MuJoCo.
Machine Learning for Robotics: Utilizing CMU 11-785 and 10-601 foundations to integrate deep learning into the robotics stack, specifically for perception-based navigation and reinforcement learning (DAgger).
Industrial Automation: Overhauled FANUC ArcMate workflows using KAREL and ROS-I, achieving over 90 percent repeatability and a 90 percent reduction in setup time.

Find my complete CV here .

Projects

Lunar ROADSTER

MRSD Capstone Project, Carnegie Mellon University | Lunar Robotic Operator for Autonomous Development of Surface Trails and Exploration Routes

Mars Rover

Mars Rover Manipal, Manipal Institute of Technology

Automated Additive Manufacturing

Manufacturing Futures Institute, Carnegie Mellon University

Dexterous Bimanual and Humanoid Manipulation

Adapting Object-Centric Policies from Dual-Arm Systems to Humanoid Embodiments

OpenCR Learning Kit

Continuum Robotics Laboratory, University of Toronto

Coffee Barista

A robot arm is tasked with pouring beads representing different coffee ingredients

Publications

2024

SmartCradle

SmartCare Baby: Revolutionizing Infant Safety and Comfort with the Smart Auto-Cleaning Cradle System

Srishti Gupta, Ankit Aggarwal, and Narendra Khatri

In 2024 4th International Conference on Technological Advancements in Computational Sciences (ICTACS), 2024

Abs DOI

Nowadays, in metropolises, both parents are working; therefore, it is of paramount importance to ensure the well-being of the infant. As a result, it is challenging to ensure the infant’s safety, security, and comfort. The present study introduces an innovative solution that uses humidity and sound sensors for continuous monitoring of the infant’s bed and environmental conditions in the cradle. Along with that, a camera is also installed for continuous monitoring using live streaming through the live mobile application. The notable new addition to this study was the integration of the auto-cleaning system into the smart cradle, which is yet unexplored in the field of smart cradle systems for infant care. The Raspberry Pi control was used as the central control for the monitoring system, and Arduino was deployed for the control of the cradle movement, which used the PWM-based control of the motor for the cradle moment. As infants urinate and start crying, it is detected through a sound sensor, and it starts shaking the cradle while simultaneously checking the moisture conditions in the bed, which updates the parent with a message on their mobile. On the request of the parent, the auto-cleaning system can be used for the change of the sheet, which also ensures the safety of the infant by taking feedback using an ultrasonic sensor that the baby is not in the cradle and then changing the sheet using the motorized control. This illustrates the potential of the IoT in infant care and safety. By actively monitoring sound and humidity levels, caregivers gain valuable insights into the infant’s needs, allowing them to make real-time adjustments to the cradle’s motions. Incorporating a camera and a mobile app further enhances monitoring and accessibility.
3-RPR

Workspace and Trajectory Analysis of the 3-RPR Planar Parallel Manipulators with Joint Clearance

Ankur Jaiswal, Sidhant Barai, Abhishek Jha, and 2 more authors

In Recent Advances in Machines, Mechanisms, Materials and Design, 2024

Abs DOI

The parallel mechanisms are mostly used in industrial and robotics applications for achieving desired tasks. The performance of the mechanism depends upon the many parameters like link joint clearance and manufacturing defects. A new mathematical model is developed for the analysis of linear input (prismatic joint) mechanisms under the effects of joint clearance. This paper presents the loop closure equation used for estimating the mechanical/position error due to the effect of link tolerances. The results evaluate the position error using a mathematical approach. The results are validated by the CAD model approach. The demonstrated case is illustrated in this paper.