Dynamic requirements for stability and maneuverability during locomotion

        The mechanics and control of stability and maneuverability are little understood compared to constant-average-speed locomotion. Several studies have focused on maintaining stability during standing or locomotion, with a focus on preventing injury (Cham and Redfern, 2002; Maki et al., 2003; Patla, 2003), and some research has focused on postural control using a variety of methods, from direct perturbations to dynamical systems analysis of naturally-occurring movement variability (Allum et al., 2003; Collins and De Luca, 1993; Horak and Nashner, 1986; Loughlin et al., 2003; Nashner and Forssberg, 1986; Winter, 1995). Biomechanical models have served to generate hypotheses for control laws at different levels of musculoskeletal organization (Gage et al., 2004; Khang and Zajac, 1989a; Khang and Zajac, 1989b; Menegaldo et al., 2003; Morasso and Schieppati, 1999; Park et al., 2004; Winter et al., 2003). Recently, direct perturbations, and analyses of naturally-occurring variability have also been used to study locomotory stabilization (Eng et al., 1997; Hurmuzlu and Basdogan, 1994; Mackinnon and Winter, 1993; Misiaszek et al., 2000; Oddsson et al., 2004). As in the study of posture, biomechanical models have served to understand mechanics and control. Sensitivity analyses situate the range of observed responses within the set of mechanically successful strategies (Patton et al., 1999; Yang et al., 1990), and can reveal passive and active contributions to stabilization during locomotion (Bauby and Kuo, 2000; Donelan et al., 2004; Gunther et al., 2004; Lyon and Day, 1997; McGeer, 1990a; McGeer, 1990b; Seyfarth et al., 2002; Seyfarth et al., 2003). However, few studies have directly linked motor control strategies used for locomotory stabilization with mechanical outcomes.

         Maneuvering presents an ideal system in which neural control of muscle function can be directly linked to mechanical outcomes. Moreover, maneuvering is a repeatable perturbation that can be used to understand the neural control mechanisms underlying stabilization of locomotion. Fundamental questions about the mechanics of maneuverability remain unanswered. For example, the relationships between maneuverability, anthropometry, and muscular physiology are largely unexplored.
 
        We hypothesize that two mechanical requirements constrain the mechanisms available for maneuvering in the horizontal plane (i.e. turning): A) The movement direction of the center of mass (COM) must be changed, and B) the body must be rotated to face the new direction of COM movement. A force impulse must deflect the COM movement direction, and an appropriate torque impulse must also be generated to rotate the body. Based on these force and torque requirements, we developed a simple mathematical model of the task requirements for maneuvering that made specific predictions of the force production necessary to turn during running. We initially tested the model using insects and found that it was able to describe leg force production for each of the legs. In collaboration with Thor Besier and David Lloyd (at Stanford and University of Western Australia, respectively), we tested the model on human cutting maneuvers during running. We found that this simple algebraic model could explain 70% of the variance in braking forces during sidestep cutting maneuvers of different degrees, and surprisingly even of crossover cutting maneuvers . Braking forces observed during running turns contributed to prevent body over-rotation during turning. Recent experiments in collaboration with Alan Wilson at the Royal Veterinary College (U.K.) have shown that the simple model can also explain the ground-reaction forces used by ostriches to turn. Based on morphology and behavior, ostriches were hypothesized to require less braking force than humans during running turns, which was observed.