The intentional spring: a strategy for modeling systems that learn to perform intentional acts.

In motor task learning by instruction, the instructor's skill and intention, which, initially, are extrinsic constraints on the learner's perceiving and acting, eventually become internalized as intrinsic constraints by the learner. How is this process to be described formally? This process takes place via a forcing function that acts both as an anticipatory (informing) influence and a hereditary (controlling) influence. A mathematical strategy is suggested by which such intentions and skills might be dynamically learned. A hypothetical task is discussed in which a blindfolded learner is motorically instructed to pull a spring to a specific target in a specific manner. The modeling strategy involves generalizing Hooke's law to the coupled instructor-spring-Learner system. Specifically, dual Volterra functions express the anticipatory and hereditary influences passed via an instructor-controlled forcing function on the shared spring. Boundary conditions (task goals) on the instructor-spring system, construed as a mathematical (self-adjoint) operator, are passed to the learner-spring system. Psychological interpretation is given to the involved mathematical operations that are passed, and mathematical (Hilbert-Schmidt's and Green's function) techniques are used to account for the release of the boundary conditions by the instructor and their absorption by the learner, and an appropriate change of their power spectra.

The acceleration-velocity relationship: identification of normal and spastic upper extremity movement.

This study defines and evaluates a non-invasive technique that will identify a spasm as an uncommanded movement (U) from a normally commanded movement (C). Upper arm function during a tracking task is recognized as non-linear and characterized by a phase plane with acceleration (second derivative of stick position) plotted on the ordinate and velocity (first derivative of stick position) plotted on the abscissa. An acceleration time constant (tau A) is defined as (VA/AM) where AM is the maximal acceleration and VA is the velocity at AM. A deceleration time constant (tau D) is also defined as (VD/DM), where DM is the maximal deceleration and VD is the velocity at DM. Ten spastic subjects (S) were categorized into three groups of mildly spastic (M), intermediately spastic (I), and highly spastic (H) which were compared to five normal subjects (N). The results indicate that for C and U, both VA and AM are significantly lower for S at all three levels as compared to N. tau A is significantly higher for M and I as compared to N. Both VD and tau D are significantly higher for S at all three levels when compared to N. DM is significantly lower for S at all three levels of S when compared to N. Finally, AM and tau A significantly differentiate C vs U for M. VA and tau D significantly differentiate for I. VA, AM as well as DM, tau D significantly differentiate for H. The physical significance of these results is discussed with respect to the time course and magnitude of net muscle moment (M), and the reversal of the agonist-antagonist relationship during the acceleration and deceleration phases. Finally, a spasm identification algorithm is proposed.

Biomimetic model of skeletal muscle isometric contraction: I. an energetic-viscoelastic model for the skeletal muscle isometric force twitch.

This paper describes a revision of the Hill-type muscle model so that it will describe the chemo-mechanical energy conversion process (energetic) and the internal-element stiffness variation (viscoelastic) during a skeletal muscle isometric force twitch contraction. The derivation of this energetic-viscoelastic model is described by a first-order linear ordinary differential equation with constant energetic and viscoelastic coefficients. The model has been implemented as part of a biomimetic model, which describes the excitation-contraction coupling necessary to drive the energetic-viscoelastic model. Finally, the energetic-viscoelastic model is validated by comparing its isometric force-time profile with that of various muscles reported in the literature.

Biomimetic model of skeletal muscle isometric contraction: II. A phenomenological model of the skeletal muscle excitation-contraction coupling process.

This paper describes a new macroscopic, phenomenological model of the skeletal muscle excitation-contraction coupling process, as represented by four principal and consecutive compartments (biophysical, biochemical, and biomechanical phases) characteristic of isometric excitation-contraction coupling in mammalian skeletal muscle, and coupled by a system of simultaneous, first-order linear ordinary differential equations. The model is based upon biological compartmental transport kinetics and irreversible thermodynamic energy transformation, and represents a distinct improvement over other biomimetic models. The model was derived using physiological parameter data published in the literature, and validated using MATLAB R12.

Microprocessor based spatial TENS (transcutaneous electric nerve stimulator) designed with waveform optimality for clinical evaluation in a pain study.

A microprocessor based TENS device is developed which utilizes a spatial procedure in the administration of electrical fields to actively interfere with pain signals reaching the brain. This unusual design also has the advantage of requiring the frequency characteristics of the electrical waveform produced to be optimally tuned to the mechanical impedance properties of the skin/tissue. Hence a much more efficient procedure for the transfer of electrical energy from the TENS device to the human tissue is provided. Data are presented involving patients from the Dayton VA Medical Center in Ohio, USA being tested with this new microprocessor system compared to the treatment obtained via a traditional stimulator.

Adaptive displays and controllers using alternative feedback.

Investigations on the design of haptic (force reflecting joystick or force display) controllers were conducted by viewing the display of force information within the context of several different paradigms. First, using analogies from electrical and mechanical systems, certain schemes of the haptic interface were hypothesized which may improve the human-machine interaction with respect to various criteria. A discussion is given on how this interaction benefits the electrical and mechanical system. To generalize this concept to the design of human-machine interfaces, three studies with haptic mechanisms were then synthesized and analyzed.

Mass discrimination under Gz acceleration.

The purpose of this study was to assess how the perception of mass discrimination is affected by elevated Gz acceleration. Previous experiments studied mass discrimination under weightless conditions. Ten subjects were tested with the Dynamic Environment Simulator (DES) at Wright-Patterson Air Force Base. Masses of 105, 110, 115, 120, and 125 g were compared to a 100-g standard for delta Ms of 5, 10, 15, 20, and 25 g. The subject had to choose which mass felt heavier. This was done at 1, 2, and 4 Gz. Significant differences were found between each of the G levels, and the subjects made more errors at higher Gz. Significant differences were also found between each of the delta Ms, except between delta Ms of 20 and 25 g. Using regression lines, the difference limen was calculated at the 75% correct response level for each Gz. The Weber fraction was found by dividing the difference limen by the 100-g standard. Weber fraction of 0.085, 0.116, and 0.145 were found at 1, 2, and 4 Gz, respectively. Impairment to discrimination was shown by calculating the ratio of the Weber fraction of the elevated Gz to 1 Gz. This demonstrated an impairment to mass discrimination at 1.36 at 2 Gz and 1.71 at 4 Gz. Impairment of mass discrimination under elevated G indicates that loss of adaptation is more important than weight or mass constancy or any other factors which would increase gravitational sensory cues. This study attempted to show adaptation by comparing runs done on different days. To show aftereffect, intervals of 1 G were compared to each other. The study did not find any adaptation or aftereffect. When compared to previous studies done in weightlessness, microgravity was found to be more detrimental to mass discrimination than macrogravity, at least up to 4 Gz.

Human tracking performance changes during combined +Gz and +/- Gy stress.

An experiment has been conducted on the centrifuge to examine roll-axis tracking performance in both static and stress environments. The stress environments were +5 Gz from 95 s and combined +5 Gz/+/- 1 or +/- 2 Gy for 95 s. Compared to the static condition, performance decrements of 19% were measured at +5 Gz, 45% at combined +5 Gz/+/- 1 Gy, and 70% at combined +5 Gz/+/-2 Gy. Heart rate increases were noted during the stress environments but no significant heart rate differences were noted between the +5 Gz and the combined +5 Gz/+/-2 Gy conditions. The conventional lap belt and shoulder harness restraint system, while not optimum, was adequate for these G environments.

Quantification of reaction time and time perception during Space Shuttle operations.

A microprocessor-based test battery containing simple reaction time, choice reaction time, and time perception tasks was flown aboard a 1985 Space Shuttle flight. Data were obtained from four crewmembers. Individual subject means indicate a correlation between change in reaction time during the flight and the presence of space motion sickness symptoms. The time perception task results indicate that the shortest duration task time (2 s) is progressively overestimated as the mission proceeds and is statistically significant (p less than 0.01) when comparing preflight and postflight baselines. The tasks that required longer periods of time to estimate (8, 12, and 16 s) are less affected.

Design of a computerized device to measure simple reaction time/decision time.

The construction and validation of a computerized device to test simple reaction time and decision-making time is discussed. The device is required to demonstrate repeatability (consistency of response times from human subjects) across replications of the same experimental condition. A second requirement of the device is that it must separately and independently measure simple reaction time and the time it takes to process one bit of information. The concepts behind the construction of the box are discussed, as well as the design of its internal circuits to achieve these specifications.

Performance study involving a force-reflecting joystick for spastic individuals performing two types of tracking tasks.

10 upper-extremity spastic subjects and 10 normal subjects were studied with a force-reflecting joystick in the performance of a continuous time-tracking task as well as an acquisition task termed, "Fitts' Law." Certain force-reflection paradigms, in a spatial sense, allowed the spastic subjects to obtain performance proficiency near levels of the normal subjects as measured by a capacity metric.

Effect of directional response variables on eye-hand reaction times and decision time.

To determine if direction of response affects reaction time, we measured the time for hand response to a visual stimulus, using a sensitive, microprocessor-based testing device to determine simple reaction time (RT), choice RT, and decision time. Mean simple RT was 207 +/- 3.7 msec. (mean +/- SEM); mean choice RT was 268 +/- 4.2 msec; and mean decision time was 61 msec. No differences were noted for leftward versus rightward movements, or midline versus lateral movements. Choice RT increased by 1.49 msec./yr. of age. Simple RT increased significantly with age for the nondominant hand, but not for the dominant hand. Right-handed subjects were more rapid with the dominant hand for choice RT. We conclude that dominance of hand tested and test initiation mechanism have major effects, but direction of movement in the lateral plane has little effect on reaction time.

Effects of haptic feedback and turbulence on landing performance using an immersive cave automatic virtual environment (CAVE).

An investigation was conducted in which subjects had to land a simulated F-16 aircraft using a CAVE (Cave Automatic Virtual Environment) facility. This was a three-dimensional virtual setting consisting of multiple mirrors, 3-D video-projected displays in a highly stressful environment employing a haptic joystick. 6 subjects learned a task which required landing in wind turbulence with a reduced visual scene. Analyses indicated that during landing, performance error variables which occurred in the same direction as the haptic forces were significantly reduced. This was true, especially when the visual scene was occluded and more reliance on the proprioceptive condition was beneficial.

Design of a haptic stick interface as a pilot's assistant in a high turbulence task environment.

A study involving 8 Air Force pilots was conducted to examine the efficacy of a force-reflecting joystick to improve performance during a simulated landing task in wind turbulence. By adding certain force characteristics to a joystick, it was of interest to see if performance may change, different control effort may be utilized, and workload measures may be altered based on the joystick utilized. The main results show that certain performance measures significantly improved by having the force reflection condition on. The implications of this study are that in certain types of precision tracking tasks, subjected to external disturbances, the addition of the force characteristics to the joystick can significantly improve performance, result in less effort for control, and lower subjective workload.

Haptic devices as a paradigm to enhance learning/function--theory and empirical studies.

Haptics is a modern term employed to describe both afferent and efferent flows of forces with humans. The appropriate design of these force flow fields when coordinated with information to and from computers and other devices to physically challenged individuals can empower the performance of certain tasks previously not attainable. Extensive empirical work presented herein has demonstrated that proper augmentation of particular dynamic force fields, synchronised with certain cognitive tasks, can lead to improved functionality. More specifically, studies in spasticity, weaken responses and having delayed human response can benefit from these exposures to force fields applied suitably. From a theoretical perspective, the use of the well-known Fitts' law provides a unifying framework from which one can quantify the benefits derived from the proper exposure to force fields. Another area where haptics offers an exciting new platform to transfer information to and from humans includes subliminal haptics. This provides a paradigm to improve situational awareness and enhances control capability, where it may be missing previously. This article updates how haptics can be used in a variety of empirical settings to enhance learning, modulate human response to assist task completion, present a theoretical framework and improve the self-esteem of physically challenged individuals.

Characterisation of a pneumatic muscle test station with two dynamic plants in cascade.

Pneumatic muscle actuators (PMAs) offer significant advantages over more traditional actuators, which make them prime candidates in rehabilitation devices. A dynamic test station (DTS) is modified to demonstrate the use of a PMA for this application. The DTS includes two dynamic systems: a PMA and a DC servomotor. An overall transfer function was developed utilising characterisation data for the PMA and DC servomotor. A Tustin (bilinear) transform was performed on the overall transfer function to obtain a discrete time system. Model parameters were optimised and used to generate input voltage profiles that achieve isokinetic (constant velocity) task specifications. Percent root mean square error values (PRMSE) between the actual and ideal profiles were used to evaluate the accuracy of this method in achieving isokinetic displacement. For PMA pressures (in kPa) of 150, 350 and 550 PRMSE were 7.80, 5.40 and 2.76, respectively.

Characterisation of a phenomenological model for commercial pneumatic muscle actuators.

This study focuses on the parameter characterisation of a three-element phenomenological model for commercially available pneumatic muscle actuators (PMAs). This model consists of a spring, damping and contractile element arranged in parallel. Data collected from static loading, contraction and relaxation experiments were fitted to theoretical solutions of the governing equation for the three-element model resulting in prediction profiles for the spring, damping and contractile force coefficient. For the spring coefficient, K N/mm, the following relationships were found: K = 32.7 - 0.0321P for 150 < or = P < or = 314 kPa and K = 17 + 0.0179P for 314 < or = P < or = 550 kPa. For the damping coefficient, B Ns/mm, the following relationship was found during contraction: B = 2.90 for 150 < or = P < or = 550 kPa. During relaxation, B = 1.57 for 150 < or = P < or = 372 kPa and B = 0.311 + 0.00338P for 372 < or = P < or = 550. The following relationship for the contractile force coefficient, F(ce) N, was also determined: F(ce) = 2.91P+44.6 for 150 < or = P < or = 550 kPa. The model was then validated by reasonably predicting the response of the PMA to a triangular wave input in pressure under a constant load on a dynamic test station.

An application of optimization methods to evaluate conditioning intervals.

A decision rule is obtained to determine when exercise is beneficial. The term beneficial used in this context refers to the extension of lifetime based on studies of cardiovascular conditioning. Models are used to characterize changes in heart rate with intense exercise. In addition, it is necessary to consider a model involving a heartbeat bank to help determine the conditioning interval. Optimization procedures are used to develop the decision rule, to examine the assumptions in the problem, and to point out when a unique solution exists. This procedure is applied to empirical data for one individual (inter-subject study) and a unique conditioning interval is obtained.

Modeling the dynamic characteristics of pneumatic muscle.

A pneumatic muscle (PM) system was studied to determine whether a three-element model could describe its dynamics. As far as the authors are aware, this model has not been used to describe the dynamics of PM. A new phenomenological model consists of a contractile (force-generating) element, spring element, and damping element in parallel. The PM system was investigated using an apparatus that allowed precise and accurate actuation pressure (P) control by a linear servo-valve. Length change of the PM was measured by a linear potentiometer. Spring and damping element functions of P were determined by a static perturbation method at several constant P values. These results indicate that at constant P, PM behaves as a spring and damper in parallel. The contractile element function of P was determined by the response to a step input in P, using values of spring and damping elements from the perturbation study. The study showed that the resulting coefficient functions of the three-element model describe the dynamic response to the step input of P accurately, indicating that the static perturbation results can be applied to the dynamic case. This model is further validated by accurately predicting the contraction response to a triangular P waveform. All three elements have pressure-dependent coefficients for pressure P in the range 207 < or = P < or = 621 kPa (30 < or = P < or = 90 psi). Studies with a step decrease in P (relaxation of the PM) indicate that the damping element coefficient is smaller during relaxation than contraction.