Underactuated Synchronous Routing

• Underactuated synchronous routing is a design strategy where fewer actuators control multiple degrees of freedom via fixed or adaptive coupling ratios.
• It leverages tendon-driven synergies and null-space-based control to synchronize motions in robotic hands and multi-agent systems.
• Adaptive synergy ensures resource-efficient, compliant operation while trading off fine, independent joint control.

Underactuated synchronous routing designates a class of mechanisms and control architectures where multiple degrees of freedom (DOF) are coordinated via fewer actuators than DOFs, while maintaining temporal or functional synchrony among outputs. In both robotic manipulation and multi-agent autonomous systems, this strategy leverages mechanical coupling, shared synergies, or priority-based control to achieve adaptive, reliable, and resource-efficient collective motion without direct actuation of every joint or subsystem.

1. Fundamental Principles

Underactuation refers to systems in which the number of independent actuators is less than the number of DOFs, resulting in configurations where not all joints or coordinates can be arbitrarily controlled. Synchronous routing specifically denotes the intentional architectural or algorithmic strategy in which a single actuation signal or mechanism causes multiple DOFs to move in sync, according to fixed or adaptively coupled ratios.

In tendon-driven robotic hands, synchronous routing manifests as all fingers (or finger joints) being actuated in parallel by a shared tendon path, transmitting mechanical input from a single actuator to multiple outputs with a fixed or synergy-inspired mapping (Yuan et al., 11 Dec 2025, Li et al., 2022, Merritt et al., 1 Apr 2026). In coordinated vehicle or agent systems, synchronous routing involves hierarchical or null-space-based algorithms, aligning the trajectories or tasks of underactuated units to maintain desired geometric relationships or group objectives (Eek et al., 2020).

2. Robotic Mechanisms: Tendon-Driven Hands

Underactuated synchronous routing is epitomized by the design of anthropomorphic robotic hands such as the UTRF-RoboHand and SoftHand Model-W.

Single-actuator synchronous transmission is achieved by routing tendons through the hand and finger structure in a manner that couples joint rotations. In the UTRF design, three phalangeal joints are mechanically coupled such that a single linear input produces proportional rotation across all joints, implemented via matched pulley/guidance radii $R_i$ and coupling tendons—yielding output angles $\theta_i = q / R_i$ for input displacement $q$ (Yuan et al., 11 Dec 2025). This fixed ratio ensures that all joints flex together, reinforcing synchronous motion.

The SoftHand Model-W extends this by routing all finger and wrist tendons through a “carpal tunnel” using Bowden sheaths, with all five-finger flexors actuated by a single spool, and all extensors by an antagonist, yielding the mapping:

$\varphi_{i,j} = x / R_p$

for all fingers $i$ and joints $j$, given common linear displacement $x$ applied by the actuator and pulley radius $R_p$ (Merritt et al., 1 Apr 2026). Contact-adaptive synergistic behavior emerges, with further motion redistributed among non-contacting joints. All routing is physically compact, minimizes distal inertia, and enables remote actuation.

3. Kinematics and Force Transmission

Exact relationships between actuator displacement, joint motion, and output forces are essential for predictable synchronous operation in underactuated systems.

For single-chain, synchronously routed fingers (UTRF), the forward kinematic equations are:

$$\theta_i = \frac{q}{R_i}, \qquad x(q) = \sum_{i=1}^3 L_i \cos\Bigl(\sum_{j=1}^i \frac{q}{R_j}\Bigr)$$

where $L_i$ is each link length (Yuan et al., 11 Dec 2025). The force balance at each joint incorporates tendon tension, joint pulley radius, and gravity. The net torque at joint $\theta_i = q / R_i$0, finger $\theta_i = q / R_i$1 in the Model-W, is:

$\theta_i = q / R_i$2

with $\theta_i = q / R_i$3 and $\theta_i = q / R_i$4 being the shared flexor and extensor tensions (Merritt et al., 1 Apr 2026).

Stiffness analysis incorporates tendon elasticity. For a tendon of stiffness $\theta_i = q / R_i$5 (Young's modulus $\theta_i = q / R_i$6, cross-section $\theta_i = q / R_i$7, length $\theta_i = q / R_i$8), the tip stiffness $\theta_i = q / R_i$9 is:

$q$0

where $q$1 is the local Jacobian and $q$2 is the diagonal tendon-stiffness matrix (Yuan et al., 11 Dec 2025).

4. Adaptive Synergies and Compliance

A cornerstone advantage of underactuated synchronous routing is “adaptive synergy”: the capacity to redistribute actuation as external constraints arise. When one finger or joint contacts an object, further actuator displacement automatically flexes the remaining joints and fingers without explicit sensing or control, enforcing conformal grasps around objects of arbitrary geometry (Merritt et al., 1 Apr 2026, Li et al., 2022).

For instance, in the Model-W, once a fingertip or phalanx contacts an object and stalls, the driving displacement is adaptively mapped into the rest of the multi-joint, multi-finger chain, ensuring enveloping, human-like manipulation with a minimal control interface (Merritt et al., 1 Apr 2026). The synergy mapping can be compactly represented as $q$3, with $q$4 the single actuation coordinate and $q$5 encoding the ratio of local displacement per joint.

Empirical testing of such hands confirms both stability and grasp adaptability in complex tasks, with measured performance (e.g., stack accuracy, joint travel, task time) correlating directly with the number and routing of shared actuators (Merritt et al., 1 Apr 2026, Yuan et al., 11 Dec 2025).

5. Multi-Agent Control: Synchronous Routing in Underactuated Vehicles

Synchronous routing generalizes to multi-agent systems, such as unmanned surface vessels (USVs) navigating in formation.

Here, “underactuated synchronous routing” is achieved via hierarchical control. The system is underactuated because each USV manipulates only surge thrust and rudder, lacking direct control over sway. Synchronization is imposed by a null-space-based (NSB) behavioral control hierarchy:

1. Collision avoidance (highest priority)
2. Maintenance of formation geometry (enforcing constant relative position)
3. Barycenter path following (causing the group to traverse a prescribed trajectory) (Eek et al., 2020)

These objectives are mapped to task-specific velocity commands via closed-loop inverse kinematics (CLIK), merged using null-space projectors to ensure task priority is never violated. The final velocity command, $q$6, is projected into the feasible subspace, maintaining synchronization and formation geometry under actuation limits.

A line-of-sight (LOS) guidance law parametrizes the barycenter’s tangent error to the path and adaptively corrects headings and speeds. The resulting closed-loop system is shown to be uniformly globally asymptotically stable (UGAS) and uniformly semiglobal exponentially stable (USGES), while the unactuated sway dynamics remain bounded.

6. Practical Implications and Performance Metrics

Demonstrated architectures show significant performance advantages attributable to synchronous underactuated routing:

• Reduced actuator count, weight, and distal inertia (Model-W: 4 motors for 7 DOFs) (Merritt et al., 1 Apr 2026)
• Adaptive, enveloping grasp capability requiring no per-joint sensors (Yuan et al., 11 Dec 2025, Li et al., 2022)
• Predictable stiffness and reliability under load, with measured/analytical agreement to within 0.5 mm deflection (UTRF) (Yuan et al., 11 Dec 2025)
• Enhanced manipulation speed and reduced arm/robot joint travel: e.g., tasks completed 29% faster and with up to 36% less arm joint excursion when the SoftHand wrist is actuated synchronously (Merritt et al., 1 Apr 2026)
• In formation path following, provable stability and disturbance rejection in multi-agent underactuated platforms (Eek et al., 2020)

7. Design Trade-Offs and Limitations

While underactuated synchronous routing confers simplicity and adaptability, certain limitations are intrinsic:

• Fine, independent joint control is not possible; all outputs move in mechanically prescribed ratios unless physical contact interrupts motion.
• The system’s compliance and adaptivity hinge on precise mechanical design (tendon routing, friction management, pulley ratios, elasticity) (Yuan et al., 11 Dec 2025, Merritt et al., 1 Apr 2026)
• For multi-agent cases, adherence to geometric synchrony is sacrificed if collision avoidance or higher-priority constraints are activated (Eek et al., 2020)

Measured performance is highly sensitive to routing geometry, actuator backlash, and tendon characteristics. Tuning preloads, stiffness, and routing paths is essential for maximizing both compliance and stability.

---

References:

• "SoftHand Model-W: A 3D-Printed, Anthropomorphic, Underactuated Robot Hand with Integrated Wrist and Carpal Tunnel" (Merritt et al., 1 Apr 2026)
• "Design and Validation of an Under-actuated Robotic Finger with Synchronous Tendon Routing" (Yuan et al., 11 Dec 2025)
• "BRL/Pisa/IIT SoftHand: A Low-cost, 3D-Printed, Underactuated, Tendon-Driven Hand with Soft and Adaptive Synergies" (Li et al., 2022)
• "Formation Path Following Control of Underactuated USVs – With Proofs" (Eek et al., 2020)