SEA vs Quasi-Direct Drive for Legged Robots — PatSnap Eureka
Series Elastic Actuator vs. Quasi-Direct Drive for Impact-Resilient Legged Robots
Two dominant paradigms — passive spring compliance versus low-ratio backdrivability — define how modern legged robots survive ground contact. Understand the mechanical, control, and performance trade-offs that determine which architecture fits your platform.
Gear Ratio Range: SEA vs QDD
SEA transmissions run 50:1–200:1; QDD systems stay at 6:1–9:1 — a 10–30× difference that defines their respective impact behaviour.
Two Philosophies, One Goal: Surviving Ground Contact
The Series Elastic Actuator (SEA) paradigm, introduced in the early 1990s, places a compliant elastic element in series between a motor-gearbox assembly and the joint output link. This architecture delivers high-fidelity force sensing via spring deflection measurement, passive shock absorption at touchdown, energy storage and return during cyclic locomotion, and inherently safe human-robot interaction. The mass-spring behaviour produced by SEA joints is directly aligned with the Spring-Loaded Inverted Pendulum (SLIP) model of running, enabling naturalistic compliant interaction with the ground.
Quasi-Direct Drive (QDD) actuators take a fundamentally different approach. Rather than inserting compliance mechanically between motor and joint, QDD systems use high-torque-density motors — typically large-diameter, low-pole-count brushless DC motors — coupled with very low gear ratios, typically in the range of 1:6 to 1:9. This configuration prioritises backdrivability, low reflected inertia, and high mechanical transparency, allowing the actuator to absorb impacts passively through motor back-EMF and the natural admittance of the drive train rather than a discrete elastic element.
Research spanning institutions including MIT, IIT Genova, Max Planck Institute, University of Michigan, Harbin Institute of Technology, and others — covering over 60 literature entries and patent records from 2011–2023 — forms the evidence base for this comparison. Understanding these distinctions is critical for R&D engineers selecting actuator architectures for next-generation legged platforms. Explore the full dataset via PatSnap Eureka.
Performance Data: SEA vs Quasi-Direct Drive
Key metrics from 60+ literature entries and patent records, spanning 2011–2023. All values sourced directly from published research.
SEA Energy Savings vs Direct Drive (Walking)
SEAs reduce energy consumption at the hip by 18–29% in walking; the eLeg hybrid achieves 65–75% electrical energy improvement (IIT Genova, 2019).
QDD Actuator Durability After 420,000 Gait Strides
University of Michigan (2022) 3D-printed QDD actuators showed only 2% efficiency reduction and 26 mrad backlash growth after 420,000 stride equivalents, built for under $200 each.
SEA vs Quasi-Direct Drive: Design Dimension Comparison
Eight engineering axes that determine which actuator architecture suits your legged robot platform, drawn from 60+ literature entries and patent records (2011–2023).
| Design Dimension | Series Elastic Actuator (SEA) | Quasi-Direct Drive (QDD) |
|---|---|---|
| Impact absorption mechanism | Passive spring deflection — decouples gearbox inertia from impulsive ground force | Low reflected inertia + backdrivability via motor back-EMF |
| Force sensing | Implicit — spring deflection provides torque estimate without external load cell | Requires external current or torque sensing |
| Force control bandwidth | Limited — spring acts as low-pass mechanical filter on force transmission | High — near-rigid transmission, current commands translate almost instantaneously |
| Energy storage | Yes — passive elastic storage reduces peak power and energy in walking by up to 29% | No — motor must supply all peak power demands |
Need to compare actuator patents across institutions?
PatSnap Eureka maps the full innovation landscape — from MIT to Max Planck to IIT Genova.
How Each Architecture Handles Impact and Force Control
The core engineering differences between SEA and QDD emerge most sharply in impact absorption, force bandwidth, and energy management — the three axes that define legged locomotion performance.
Spring Compression Decouples Gearbox from Ground
SEAs absorb impact through the deliberate compression of a physical spring element inserted in the drive train. The spring deflects during touchdown, decoupling the motor-gearbox inertia from the impulsive ground force and simultaneously providing a measurable signal for torque estimation. Compliant mechanisms in distal leg segments reduce peak impact forces compared to purely stiff structures, with nonlinear compliance profiles providing superior buffering over linear springs, as confirmed by Harbin Institute of Technology research (2017).
Passive buffering at touchdownLow Reflected Inertia via Gear Ratio Squared Reduction
QDD systems absorb impact through the low reflected inertia and backdrivability of the entire actuator. With gear ratios as low as 6:1, the reflected motor inertia is reduced by the square of the gear ratio, so the effective mechanical impedance at the joint is dominated by the link inertia rather than the motor inertia. The Max Planck Institute (2020) demonstrated a maximum dimensionless leg stiffness of 10.8 without active damping — comparable to biological legs — without any physical spring element.
Backdrivability as passive complianceElastic Element Is a Built-In Sensor — and a Bandwidth Bottleneck
SEAs use spring deflection as a built-in force sensor, enabling accurate torque estimation without external load cells. However, the elastic element acts as a low-pass mechanical filter: high-frequency force commands cannot be faithfully tracked because the spring compliance attenuates rapid changes in output force. Variable stiffness implementations address this partially but add mechanical complexity. IIT Genova (2017) notes that joint elasticity "complexifies the design of balancing and walking controllers," requiring motor velocities as intermediate control inputs for momentum regulation.
Low-pass filter on force transmissionNear-Rigid Transmission Enables High-Bandwidth Torque Control
QDD systems do not suffer the bandwidth limitation of SEAs. With a nearly rigid transmission, motor current commands translate almost instantaneously into joint torques. The Max Planck Institute (2020) open-source platform achieves low-impedance, high-bandwidth torque control at the foot entirely through current regulation without any spring element. The Université Laval PEQDD (2023) exosuit actuator benchmarks demonstrate superior control responsiveness arising from QDD's inherently low mechanical impedance and minimal compliance in the drive chain. Explore the patent landscape via PatSnap Eureka.
Current → torque, near-instantaneousWhere SEAs Excel — and the Emerging Middle Ground
SEAs possess a structural energy storage capability absent in QDD designs. By storing energy during negative-work phases and releasing it during positive-work phases, SEAs reduce both peak power and energy consumption in walking by up to 29% compared to direct drive, as quantified by TU Darmstadt (2014). As reinforced by UT Austin (2018), the performance of SEAs can actually surpass ideal force source actuators by storing and releasing energy along locomotion trajectories. This advantage is task- and speed-dependent: in running scenarios optimised purely for energy, a direct drive can outperform an SEA.
QDD systems cannot store elastic energy and therefore rely entirely on the motor to supply peak power demands, which may require larger motors or higher current ratings in jumping or impact-intensive tasks. Beijing Institute of Technology (2021) addresses this by pairing a high torque-density actuator with a custom motor and two-stage planetary — a QDD-adjacent approach — achieving 1.8 m jump heights and robust landings through nonlinear optimisation rather than passive spring energy storage.
A clear trend in the dataset is the convergence toward hybrid series-parallel elastic architectures. Harbin Institute of Technology (2022) demonstrates combined series elastic modules and parallel gas springs to satisfy both compliance and output force range requirements. IIT Genova's eLeg (2019) achieves 65–75% improvements in electrical energy consumption through series-parallel and biarticular compliant actuation. Vrije Universiteit Brussel (2013) shows that SEAs reduce motor power but not torque, motivating parallel elastic augmentation to reduce motor sizing further. For deeper patent intelligence on hybrid actuators, see how R&D teams use PatSnap to accelerate architecture decisions.
Variable stiffness SEA implementations have also been explored specifically for legged robots: the University of Michigan SiMPLeR (2021) demonstrates a mechanically simple design combining torsional springs with timing belts to achieve controllable stiffness in an antagonistic pair, validated through one-legged hopping experiments. TU Darmstadt (2017) demonstrates that variable stiffness SEAs can be tuned to antiresonance operation, yielding significant electrical energy savings by exploiting the system's natural dynamics. Access the full patent record via PatSnap Open API for programmatic analysis.
Institutions Driving SEA and QDD Research
Analysis of 60+ literature entries and patent records reveals concentrated poles of innovation across both actuator paradigms.
IIT Genova — SEA Humanoid & Biarticular Compliance
Contributes foundational work on compliant actuation, including SEA-based momentum control for humanoids and biarticular compliant leg designs. The eLeg (2019) achieves 65–75% electrical energy improvement through series-parallel and biarticular compliant actuation. IIT's 2017 work notes that joint elasticity "complexifies the design of balancing and walking controllers."
Max Planck Institute — QDD Open Torque-Controlled Quadruped
Pivotal contributor to QDD-based open torque-controlled quadruped architecture. The 2020 platform presents a brushless DC motor with low-gear-ratio transmission for a 2.2 kg quadruped, demonstrating maximum dimensionless leg stiffness of 10.8 without active damping — comparable to biological legs — without any physical spring element.
TU Darmstadt — SEA Energy Analysis & Variable Stiffness
Leads biomechanically-grounded SEA energy analysis. Quantifies energy savings of 18–29% at the hip joint during walking compared to direct drive. Demonstrates that variable stiffness SEAs can be tuned to antiresonance operation for significant electrical energy savings by exploiting the system's natural dynamics rather than fighting them.
University of Michigan — Spans Both Paradigms
Publishes QDD open-source actuator characterisation (2022) — under $200 per unit with only 2% efficiency degradation after 420,000 gait strides — and SEA variable stiffness design via SiMPLeR (2021), combining torsional springs with timing belts for controllable stiffness validated through one-legged hopping.
SEA vs QDD: Publication Activity 2011–2023
SEA-focused research accounts for the largest share of citations in the dataset. QDD emerges as a distinct and increasingly prominent alternative from 2018 onwards, particularly for high-bandwidth, impact-tolerant legged locomotion.
Relative Research Activity: SEA vs QDD Paradigms (2011–2023)
SEA research dominates the dataset across the full period; QDD emerges as a distinct paradigm from ~2018, with accelerating publication activity through 2023.
Series Elastic Actuator vs Quasi-Direct Drive — key questions answered
SEAs absorb impact through the deliberate compression of a physical spring element inserted in the drive train. The spring deflects during touchdown, decoupling the motor-gearbox inertia from the impulsive ground force and simultaneously providing a measurable signal for torque estimation. Compliant mechanisms in distal leg segments reduce peak impact forces compared to purely stiff structures, with nonlinear compliance profiles providing superior buffering over linear springs.
QDD systems absorb impact through the low reflected inertia and backdrivability of the entire actuator. With gear ratios as low as 6:1, the reflected motor inertia is reduced by the square of the gear ratio, so the effective mechanical impedance at the joint is dominated by the link inertia rather than the motor inertia. This allows the actuator to absorb impacts passively through motor back-EMF and the natural admittance of the drive train rather than a discrete elastic element.
By storing energy during negative-work phases and releasing it during positive-work phases, SEAs reduce both peak power and energy consumption in walking by up to 29% compared to direct drive. Variable stiffness SEAs can be tuned to antiresonance operation, yielding significant electrical energy savings by exploiting the system's natural dynamics. However, in running scenarios optimized purely for energy, a direct drive can outperform an SEA, revealing that the compliance benefit is task- and speed-dependent.
High gear ratio transmissions increase apparent inertia and introduce friction effects that degrade dynamic performance, particularly during contact transitions. High-gear-ratio systems can statically support loads through friction even when unpowered — a feature useful for exoskeletons but detrimental for dynamic legged locomotion requiring rapid compliant response. QDD systems, with their low gear ratios, avoid this penalty: apparent inertia remains close to the actual link inertia, enabling faster and more transparent regulation of contact forces.
SEAs use spring deflection as a built-in force sensor, enabling accurate torque estimation without external load cells. However, the elastic element acts as a low-pass mechanical filter: high-frequency force commands cannot be faithfully tracked because the spring compliance attenuates rapid changes in output force. QDD systems do not suffer this bandwidth limitation — with a nearly rigid transmission, motor current commands translate almost instantaneously into joint torques, achieving low-impedance, high-bandwidth torque control entirely through current regulation without any spring element.
QDD actuators can be fabricated as low-cost, durable systems. Research from the University of Michigan (2022) reports under $200 per actuator with only 2% efficiency degradation after 420,000 gait strides using off-the-shelf and 3D-printed components, enabling rapid iteration in legged robotics research. This contrasts with SEA designs that require additional spring, spring housing, and deflection sensor elements that increase joint complexity, mass, and cost.
Still have questions? Let PatSnap Eureka search the actuator patent literature for you.
Ask Eureka About Actuator DesignAccelerate Your Legged Robot Actuator R&D
Join 18,000+ innovators already using PatSnap Eureka to search patents, map innovation landscapes, and make faster architecture decisions.
References
- Development, Analysis, and Control of Series Elastic Actuator-Driven Robot Leg — DGIST, South Korea, 2019
- Design and Characterization of a Novel High-Power Series Elastic Actuator for a Lower Limb Robotic Orthosis — Rice University, USA, 2013
- Momentum control of humanoid robots with series elastic actuators — IIT Genova, Italy, 2017
- Energetic and Peak Power Advantages of Series Elastic Actuators in an Actuated Prosthetic Leg for Walking and Running — TU Darmstadt, Germany, 2014
- SiMPLeR: A Series-Elastic Manipulator with Passive Variable Stiffness for Legged Robots — University of Michigan, USA, 2021
- Stiffness Control of Variable Serial Elastic Actuators: Energy Efficiency through Exploitation of Natural Dynamics — TU Darmstadt, Germany, 2017
- An Open Torque-Controlled Modular Robot Architecture for Legged Locomotion Research — Max Planck Institute for Intelligent Systems, Germany, 2020
- The dynamic effect of mechanical losses of transmissions on the equation of motion of legged robots — University of Illinois at Urbana-Champaign, USA, 2021
- Design and Characterization of 3D Printed, Open-Source Actuators for Legged Locomotion — University of Michigan, USA, 2022
- Design of a Quasi-Direct Drive Actuator with Embedded Pulley for a Compact, Lightweight, and High-Bandwidth Exosuit — Université Laval, Canada, 2023
- Exploiting the Natural Dynamics of Series Elastic Robots by Actuator-Centered Sequential Linear Programming — University of Texas at Austin, USA, 2018
- On the utility of leg distal compliance for buffering landing impact of legged robots — Harbin Institute of Technology, China, 2017
- Impact Mitigation for Dynamic Legged Robots with Steel Wire Transmission Using Nonlinear Active Compliance Control — Harbin Institute of Technology, China, 2021
- Design and Implementation of Symmetric Legged Robot for Highly Dynamic Jumping and Impact Mitigation — Beijing Institute of Technology, China, 2021
- The Analysis of Mechanical Structure of a Robotic Leg in Running for Impact Mitigation — Sogang University, South Korea, 2020
- Concept of a Series-Parallel Elastic Actuator for a Powered Transtibial Prosthesis — Vrije Universiteit Brussel, Belgium, 2013
- An efficient leg with series–parallel and biarticular compliant actuation: design optimization, modeling, and control of the eLeg — IIT Genova, Italy, 2019
- Design and Control of a Series–Parallel Elastic Actuator for a Weight-Bearing Exoskeleton Robot — Harbin Institute of Technology, China, 2022
- IEEE — Institute of Electrical and Electronics Engineers (robotics and control systems standards)
- Nature — Biological locomotion and biomechanics reference literature
All data and statistics on this page are sourced from the references above and from PatSnap's proprietary innovation intelligence platform.
PatSnap Eureka searches patents and research to answer instantly.