About

Koushil Sreenath is an Assistant Professor of Mechanical Engineering at UC Berkeley. He holds a Ph.D. in Electrical Engineering and Computer Science and an M.S. in Applied Mathematics from the University of Michigan at Ann Arbor, obtained in 2011. His professional background includes postdoctoral research at the GRASP Lab at the University of Pennsylvania from 2011 to 2013 and an assistant professorship at Carnegie Mellon University from 2013 to 2017. His research interests lie at the intersection of highly dynamic robotics and applied nonlinear control, with notable work on dynamic legged locomotion on the bipedal robot MABEL and dynamic aerial manipulation, which have been featured on various media outlets including The Discovery Channel, CNN, ESPN, FOX, CBS, IEEE Spectrum, New Scientist, and Huffington Post. Additionally, he has contributed to the field through publications such as a book on adaptive sampling with mobile sensor networks. His work has earned him several awards, including the Best Paper Award at the Robotics: Science and Systems Conference in 2013 and the Google Faculty Research Award in Robotics in 2015.

Research topics

• Artificial Intelligence
• Computer Science
• Mathematical optimization
• Mathematics
• Algorithm
• Control engineering
• Engineering

Selected publications

• Learning Smooth Humanoid Locomotion through Lipschitz-Constrained Policies

  2025-10-19 · 1 citations

  article

  Reinforcement learning combined with sim-to-real transfer offers a general framework for developing locomotion controllers for legged robots. To facilitate successful deployment in the real world, smoothing techniques, such as low-pass filters and smoothness rewards, are often employed to develop policies with smooth behaviors. However, because these techniques are non-differentiable and usually require tedious tuning of a large set of hyperparameters, they tend to require extensive manual tuning for each robotic platform. To address this challenge and establish a general technique for enforcing smooth behaviors, we propose a simple and effective method that imposes a Lipschitz constraint on a learned policy, which we refer to as Lipschitz-Constrained Policies (LCP). We show that the Lipschitz constraint can be implemented in the form of a gradient penalty, which provides a differentiable objective that can be easily incorporated with automatic differentiation frameworks. We demonstrate that LCP effectively replaces the need for smoothing rewards or low-pass filters and can be easily integrated into training frameworks for many distinct humanoid robots. We extensively evaluate LCP in both simulation and real-world humanoid robots, producing smooth and robust locomotion controllers. All simulation and deployment code, along with complete checkpoints, is available on our project page: https://lipschitz-constrained-policy.github.io.

  PublisherDOI

• Demonstrating Berkeley Humanoid Lite: An Open-source, Accessible, and Customizable 3D-printed Humanoid Robot

  2025-06-21 · 6 citations

  articleOpen accessSenior author

  Despite significant interest and advancements in humanoid robotics, most existing commercially available hardware remains high-cost, closed-source, and non-transparent within the robotics community.This lack of accessibility and customization hinders the growth of the field and the broader development of humanoid technologies.To address these challenges and promote democratization in humanoid robotics, we demonstrate Berkeley Humanoid Lite, an open-source humanoid robot designed to be accessible, customizable, and beneficial for the entire community.The core of this design is a modular 3D-printed gearbox for the actuators and robot body.All components can be sourced from widely available e-commerce platforms and fabricated using standard desktop 3D printers, keeping the total hardware cost under $5,000 (based on U.S. market prices).The design emphasizes modularity and ease of fabrication.To address the inherent limitations of 3D-printed gearboxes, such as reduced strength and durability compared to metal alternatives, we adopted a cycloidal gear design, which provides an optimal form factor in this context.Extensive testing was conducted on the 3D-printed actuators to validate their durability and alleviate concerns about the reliability of plastic components.To demonstrate the capabilities of Berkeley Humanoid Lite, we conducted a series of experiments, including the development of a locomotion controller using reinforcement learning.These experiments successfully showcased zero-shot policy transfer from simulation to hardware, highlighting the platform's suitability for research validation.By fully open-sourcing the hardware design, embedded code, and training and deployment frameworks, we aim for Berkeley Humanoid Lite to serve as a pivotal step toward democratizing the development of humanoid robotics.

  PublisherDOI

• Unraveling the Secrets of the lower Solar Atmosphere: One year of Operation of the Solar Ultraviolet Imaging Telescope (SUIT) on board Aditya-L1

  ArXiv.org · 2025-08-12

  preprintOpen access

  The Solar Ultraviolet Imaging Telescope (SUIT) is an instrument onboard Aditya--L1, the first solar space observatory of the Indian Space Research Organization (ISRO), India, launched on September 2, 2023. SUIT is designed to image the Sun in the 200--400 nm wavelength band in eight narrowband and three broadband filters. SUIT's science goals start with observing the solar atmosphere and large-scale continuum variations, the physics of solar flares in the NUV region, and many more. The paper elucidates the functioning of the instrument, software packages developed for easier calibration, analysis, and feedback, calibration routines, and the regular maintenance activity of SUIT during the first year of its operation. The paper also presents the various operations undergone by, numerous program sequences orchestrated to achieve the science requirements, and highlights some remarkable observations made during the first year of observations with SUIT.

  PublisherOA PDFDOI

• Codesign of a Differential Three-Segment Leg Enhancing Payload and Efficiency in Quadrupeds

  IEEE/ASME Transactions on Mechatronics · 2025-08-19

  article

  Roboticsrequires innovative design and system-level optimization to enhance performance. This study presents a differential three-segment leg and a codesign optimization framework aimed at improving the load capacity and energy efficiency of quadrupeds. The proposed leg design reduces the actuator torque for the same ground reaction force, decreasing the actuator Joule heating. The codesign framework simultaneously optimizes actuator, leg mechanism, and locomotion variables to minimize energy consumption. A hierarchical penalty function is used to address design constraints, such as leg workspace and actuator speed limits, transforming the constrained problem into an unconstrained one. Using the differential evolution algorithm, optimization is achieved efficiently, with convergence in 30 s. Based on these results, we developed the quadrupedal robot: Sirius-plus. Experimental validation, including single-leg swing tests and load-bearing walking trials, demonstrated the enhanced payload capacity and energy efficiency of our design.

  PublisherDOI

• CHyLL: Learning Continuous Neural Representations of Hybrid Systems

  ArXiv.org · 2025-12-10

  preprintOpen accessSenior author

  Learning the flows of hybrid systems that have both continuous and discrete time dynamics is challenging. The existing method learns the dynamics in each discrete mode, which suffers from the combination of mode switching and discontinuities in the flows. In this work, we propose CHyLL (Continuous Hybrid System Learning in Latent Space), which learns a continuous neural representation of a hybrid system without trajectory segmentation, event functions, or mode switching. The key insight of CHyLL is that the reset map glues the state space at the guard surface, reformulating the state space as a piecewise smooth quotient manifold where the flow becomes spatially continuous. Building upon these insights and the embedding theorems grounded in differential topology, CHyLL concurrently learns a singularity-free neural embedding in a higher-dimensional space and the continuous flow in it. We showcase that CHyLL can accurately predict the flow of hybrid systems with superior accuracy and identify the topological invariants of the hybrid systems. Finally, we apply CHyLL to the stochastic optimal control problem.

  PublisherOA PDFDOI

• Demonstrating MuJoCo Playground

  2025-06-21 · 3 citations

  articleOpen access

  We introduce MuJoCo Playground, a fully opensource framework for robot learning built with MJX, with the express goal of streamlining simulation, training, and simto-real transfer onto robots.With a simple pip install playground, researchers can train policies in minutes on a single GPU.Playground supports diverse robotic platforms, including quadrupeds, humanoids, dexterous hands, and robotic arms, enabling zero-shot sim-to-real transfer from both state and pixel inputs.This is achieved through an integrated stack comprising a physics engine, batch renderer, and training environments.Along with video results, the entire framework is freely available at playground.mujoco.org.

  PublisherDOI

• Coordinated Humanoid Manipulation with Choice Policies

  ArXiv.org · 2025-12-31

  articleOpen access

  Humanoid robots hold great promise for operating in human-centric environments, yet achieving robust whole-body coordination across the head, hands, and legs remains a major challenge. We present a system that combines a modular teleoperation interface with a scalable learning framework to address this problem. Our teleoperation design decomposes humanoid control into intuitive submodules, which include hand-eye coordination, grasp primitives, arm end-effector tracking, and locomotion. This modularity allows us to collect high-quality demonstrations efficiently. Building on this, we introduce Choice Policy, an imitation learning approach that generates multiple candidate actions and learns to score them. This architecture enables both fast inference and effective modeling of multimodal behaviors. We validate our approach on two real-world tasks: dishwasher loading and whole-body loco-manipulation for whiteboard wiping. Experiments show that Choice Policy significantly outperforms diffusion policies and standard behavior cloning. Furthermore, our results indicate that hand-eye coordination is critical for success in long-horizon tasks. Our work demonstrates a practical path toward scalable data collection and learning for coordinated humanoid manipulation in unstructured environments.

  PublisherOA PDF

• Estimation of Aerodynamics Forces in Dynamic Morphing Wing Flight

  2025-10-19

  article

  Accurate estimation of aerodynamic forces is essential for advancing the control, modeling, and design of flapping-wing aerial robots with dynamic morphing capabilities. In this paper, we investigate two distinct methodologies for force estimation on Aerobat, a bio-inspired flapping-wing platform designed to emulate the inertial and aerodynamic behaviors observed in bat flight. Our goal is to quantify aerodynamic force contributions during tethered flight, a crucial step toward closed-loop flight control. The first method is a physics-based observer derived from Hamiltonian mechanics that leverages the concept of conjugate momentum to infer external aerodynamic forces acting on the robot. This observer builds on the system’s reduced-order dynamic model and utilizes real-time sensor data to estimate forces without requiring training data. The second method employs a neural network-based regression model, specifically a multi-layer perceptron (MLP), to learn a mapping from joint kinematics, flapping frequency, and environmental parameters to aerodynamic force outputs. We evaluate both estimators using a 6-axis load cell in a high-frequency data acquisition setup that enables fine-grained force measurements during periodic wingbeats. The conjugate momentum observer and the regression model demonstrate strong agreement across three force components (F<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</inf>, F<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">y</inf>, F<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">z</inf>).

  PublisherDOI

• Adaptive Energy Regularization for Autonomous Gait Transition and Energy-Efficient Quadruped Locomotion

  2025-05-19 · 7 citations

  article

  In reinforcement learning for legged robot locomotion, crafting effective reward strategies is crucial. Predefined gait patterns and complex reward systems are widely used to stabilize policy training. Drawing from the natural locomotion behaviors of humans and animals, which adapt their gaits to minimize energy consumption, we investigate the impact of incorporating an energy-efficient reward term that prioritizes distance-averaged energy consumption into the reinforcement learning framework. Our findings demonstrate that this simple addition enables quadruped robots to autonomously select appropriate gaits-such as four-beat walking at lower speeds and trotting at higher speeds-without the need for explicit gait regularizations. Furthermore, we provide a guideline for tuning the weight of this energy-efficient reward, facilitating its application in real-world scenarios. The effectiveness of our approach is validated through simulations and on a real Unitree Gol robot. This research highlights the potential of energy-centric reward functions to simplify and enhance the learning of adaptive and efficient locomotion in quadruped robots. Videos and more details are at https://sites.google.com/berkeley.edu/efficient-locomotion

  PublisherDOI

• Learning Dexterous Manipulation Skills from Imperfect Simulations

  arXiv (Cornell University) · 2025-12-01

  preprintOpen access

  Reinforcement learning and sim-to-real transfer have made significant progress in dexterous manipulation. However, progress remains limited by the difficulty of simulating complex contact dynamics and multisensory signals, especially tactile feedback. In this work, we propose \ours, a sim-to-real framework that addresses these limitations and demonstrates its effectiveness on nut-bolt fastening and screwdriving with multi-fingered hands. The framework has three stages. First, we train reinforcement learning policies in simulation using simplified object models that lead to the emergence of correct finger gaits. We then use the learned policy as a skill primitive within a teleoperation system to collect real-world demonstrations that contain tactile and proprioceptive information. Finally, we train a behavior cloning policy that incorporates tactile sensing and show that it generalizes to nuts and screwdrivers with diverse geometries. Experiments across both tasks show high task progress ratios compared to direct sim-to-real transfer and robust performance even on unseen object shapes and under external perturbations. Videos and code are available on https://dexscrew.github.io.

  PublisherOA PDFDOI

Recent grants

• NRI: Collaborative Research: Unified Feedback Control and Mechanical Design for Robotic, Prosthetic, and Exoskeleton Locomotion

  NSF · $182k · 2018–2020

• Geometric Control for Dynamic Aerial Manipulation and Transportation

  NSF · $109k · 2018–2019

• CRII: RI: Dynamic Multi-Robot Coordination and Cooperation Using Dynamically Stable Mobile Robots

  NSF · $168k · 2015–2017

• Geometric Control for Dynamic Aerial Manipulation and Transportation

  NSF · $357k · 2015–2018

• CAREER: Control and Fractal-based Stability of Dynamic Vision-Based Aperiodic Legged Locomotion

  NSF · $586k · 2020–2026

Frequent coauthors

• Zhongyu Li

  Southern Medical University

  59 shared

• Jun Zeng

  Yunnan University of Finance And Economics

  36 shared

• Ayush Agrawal

  Jaguar Land Rover (United Kingdom)

  35 shared

• Bike Zhang

  31 shared

• Claire J. Tomlin

  25 shared

• Jason J. Choi

  22 shared

• Jessy W. Grizzle

  University of Michigan–Ann Arbor

  18 shared

• Quan Nguyen

  17 shared

Education

• Ph.D., Electrical Engineering and Computer Science

  University of Michigan at Ann Arbor

  2011

• M.S., Applied Mathematics

  University of Michigan at Ann Arbor

  2011

Awards & honors

• Google Faculty Research Award – Robotics (2015)
• NSF CISE Research Initiation Initiative Award (2015)
• Marquis Who’s Who in America (2015)
• Donald L. and Rhonda Struminger Faculty Fellow – Mechanical…
• Best Paper Award – Robotics: Science and Systems (RSS) (2013…