Abstract
This study investigates whether human arm movements can be explained by a "Final Position Control" (FPC) strategy—a method where the central nervous system simply sets muscle activations to a target equilibrium and allows the limb's intrinsic mechanical properties to guide it to the destination. Using a mathematical simulation of a two-link planar arm, the research demonstrates that while FPC can produce movement, it fails to replicate the straight paths and smooth, bell-shaped velocity profiles observed in humans. These findings suggest that the brain must actively plan the trajectory of a movement rather than just its final position.
The Problem of Multi-Joint Control
Controlling a multi-jointed limb like the human arm is a significant computational challenge due to nonlinear dynamics. Muscles possess complex spring-like and viscous properties, and moving one joint (e.g., the shoulder) generates "interaction torques" that affect other joints (e.g., the elbow).
A long-standing hypothesis in motor control was Final Position Control (FPC). This theory posits that the nervous system avoids solving complex differential equations in real-time by exploiting the spring-like nature of muscles. By simply picking the muscle activation levels required to hold the arm steady at a target location, the arm would naturally "snap" to the new equilibrium. This study tests whether this computationally efficient "set and forget" approach is physically viable for producing human-like motion.
Methodology
The author developed a mathematical simulation of a human arm moving in a horizontal plane, effectively eliminating gravity to isolate the dynamics of the limb itself.
The Model
The simulation utilized a two-link planar manipulator (shoulder and elbow) powered by six muscles:
- Two antagonistic pairs (one for each joint).
- One pair of "two-joint" muscles (e.g., biceps and triceps) crossing both the shoulder and elbow.
Muscles were modeled as variable-stiffness springs with linear viscosity and a 25ms first-order time lag to simulate force production delays. Biomechanical parameters were derived from anthropomorphic data and previous studies on arm dynamics.
Control Strategies
Four variations of FPC were tested:
- Pure FPC (pFPC): Specifies only the final equilibrium position.
- Directed Acceleration (dFPC): Adds a constraint ensuring the initial force is directed toward the target.
- Directed Stiffness (sFPC): Shapes the potential energy field around the target to penalize path deviations.
- Synergical FPC (cFPC): Groups muscles into functional units to simplify the control signal.
The simulation was implemented in Lisp and executed on MIT Lisp Machines using a fourth-order Runge-Kutta algorithm for integration.
Key Insights and Results
The simulations revealed that simple equilibrium-point control is insufficient for generating natural movement.
The "Wild" Nature of Simple Control
Under the Pure FPC strategy, the arm's path was often highly curved and "wild." Without active management of the trajectory, the mechanical interactions between the shoulder and elbow caused the hand to veer significantly off-course before eventually settling at the target.
The Cost of Straightness
While the Directed Stiffness (sFPC) strategy produced the straightest trajectories, it required extremely high levels of "co-contraction" (tensing all muscles simultaneously). These stiffness levels were an order of magnitude higher than those typically observed in human limbs, making the strategy biologically unrealistic.
Velocity Mismatch
Human movements typically exhibit a symmetrical, bell-shaped velocity profile. None of the FPC strategies could reliably replicate this feature; instead, they produced asymmetrical and jagged velocity profiles that did not match experimental data.
Discussion and Conclusion
The results of this study indicate that "single command" schemes like FPC are qualitatively inadequate for explaining human arm movements. While FPC is an "open-loop" strategy that might be useful in cases where sensory feedback is lost (deafferentation), it lacks the flexibility and precision required for normal, graceful motion.
The author concludes that the brain likely employs a hybrid approach: a high-level trajectory plan that dictates the specific path of the hand, superimposed on a lower-level equilibrium-point controller that utilizes the spring-like properties of the muscles. This suggests that the nervous system does indeed perform the complex task of path planning, rather than relying on a simple "set and forget" mechanism.
Summary updated