· · · Institutional Repository

Body powered prostheses have many advantages: They are reliable, lightweight and relatively cheap. The disadvantage is the need of a shoulder harness, which causes discomfort, pain and trouble donning and doffing the prosthesis. The goal of this thesis is to develop a body-powered prosthesis without the need for a shoulder harness. This is realized by making a design that uses passive wrist flexion of the prosthesis itself to operate the grasping mechanism. The force and displacement are converted to a grasping motion by using a hydraulic system. The grasping force is enhanced by a pressure intensifier and holding an object is achieved by including an automatic lock.

The number of degrees of freedom (DOF’s) required for the task of opening and closing a door by a service robot has not been a subject of research. This paper analyses the implications of using nonholonomic non-redundant service robots for this task as part of our research studying how the reduction of DOF's effects a robot's capability. We will show only one DOF remains after the door handle has been grabbed by a robot. Using mathematical simulations the dependence of the robot's path on the geometry of the door-robot system is analysed. As a result we find the system shows both stable and unstable behavior. Analysing the influence of walls and system geometry on the required positioning accuracy of a service robot we present a method to determine the robot dimensions. The width of the robot's base and the length of the robot's arm are the most important parameters.

Keeping the torso of a running robot upright is a complex task, as the torso is subject to a multitude of forces throughout the running cycle. Often a PD-control scheme is used to keep the torso in the desired upright orientation. Recently, a new torso control scheme was introduced, called Virtual Pivot Point (VPP) control. The VPP-controller stabilises the torso by applying a hip torque such that the Ground Reaction Force (GRF) is directed at a fixed point on the torso. The objective of this simulation study is to compare the performance of the VPP-controller against the PD-controller. Performance is quantified by the maximum allowable disturbance for which the simulation model can continue running. Computationally demanding simulations are performed on a simple running model and accurate results are acquired on a realistic running model. Results show that PD control outperforms VPP control in all tested situations. In theory VPP control is promising, but its implementation needs improvement.

This study quantifies and explores the nonlinearities of human arm responses to large force perturbations while subjects (n=10) performed either a position or relax task. Continuous perturbations with large variations of amplitude levels (RMS values of 2.5, 7.5, 22.5 mm displacements) and pulse perturbations with large amplitudes (average of 16 cm displacements) were applied at the hand by a 2-DOF robotic manipulator. Linear multivariable identification techniques were used to estimate the endpoint mechanical admittance from the continuous perturbations. The admittance is the relationship between force input and displacement output. Reflexive and intrinsic parameters of a 2-DOF linear arm model were fitted onto the estimated endpoint admittance. This model was used to predict the pulse perturbations. In particular, we determined to what extent human arm displacements in response to large amplitude force pulses can be predicted from identification of continuous perturbations with small amplitudes. Results showed that the estimated admittance for the relax task is a factor 18 larger compared to the estimated admittance for the position task. For the position and relax task, the estimated admittance respectively decreased with a factor 1.5 and increased with a factor 3.8 at the largest perturbation amplitude compared to the lowest amplitude. For the position task, this effect probably resulted from adaptation to the perturbation. The nonlinearity observed for the relax task might be well explained by nonlinear muscle properties such as the short range stiffness. On average, model predictions underestimated the peak displacements in response to the pulse stimulus by a factor 1.7. This shows the need to include nonlinearities in models for pulse shaped loading conditions.

Over 0.9% of the population suffers from a movement disorder. The pathophysiology of most movement disorders remains unknown, thereby impairing effective diagnosis and consequently effective treatment. Abnormal activity of the cerebellum (CBL) and basal ganglia (BG) has been implicated in many movement disorders, including Parkinsonian tremor and dystonia. Selectively activating these brain regions may help identify pathological changes and expedite diagnosis. Dedicated pairs of isometric wrist flexion tasks with and without visual feedback of the exerted torque were devised to selectively activate the CBL and BG in healthy subjects (N=5), while ensuring safety and keeping artifacts to a minimum. Increased activity in CBL and BG (putamen) was found during a constant torque task with visual feedback compared to a constant torque task without visual feedback. Increased BG (caudate nucleus) activity was found when comparing a torque task with visual feedback where flexion and rest were rapidly alternated, to the same task without visual feedback. Increased activity in the CBL was found during a constant torque task with visual feedback with an added visual error compared to a constant torque task with normal visual feedback. This study shows that specific pairs of motor tasks using the wrist and simple MR-compatible equipment allow for targeted activation of CBL and BG and paves the way for more extensive research and eventually improved diagnosis of patients.

Human walking is remarkably robust, versatile and energy-efficient: humans have the ability to handle large unexpected disturbances, perform a wide variety of gaits and consume little energy. A bipedal walking robot that performs well on all of these aspects has not yet been developed. Some robots are versatile, others are energy-efficient, and none are robust since all robots often lose balance. This lack of performance impedes their applicability in daily life. Also, it indicates that the fundamental principles of walking are not adequately understood. The goal of this thesis is to increase the understanding of the mechanics and control of bipedal locomotion and thereby increase the performance of robotic bipedal locomotion. This increased understanding will also be useful for the development of robotic devices that can help people with a decreased ambulatory ability or that can augment the performance of able-bodied persons. Bipedal locomotion is in essence about the ability to maintain control over the position and velocity of the body's center of mass (CoM). This requires controlling the forces that act on the CoM through the foot. The contact forces between the foot and the ground can be manipulated to some extent through ankle torques or upper body motions, but are mostly determined by the location of the foot relative to the CoM. The limited influence that ankle torques and upper body motions have on the contact forces and consequently on the CoM is best illustrated when one tries to remain balanced on one foot without taking a step. When slightly perturbed, balance is quickly lost and a step must be taken to prevent a fall. This demonstrates that balance control in walking relies on adequate control of foot placement (i.e., the location and timing of a step), which therefore is our main focus in the control of robotic gait. The focus on foot placement control is different from other popular control approaches in robotics. In ZMP-based control, one typically adjusts the robot's state to achieve a predefined foot placement. In Limit Cycle Walking, passive system dynamics mostly determine foot placement. This thesis presents foot placement strategies that can be adapted both in step time and step location, are an explicit function between the initial robot state and the desired future robot state, and are computationally relatively inexpensive to allow for real-time application on the robot. The contributions of this thesis to bipedal walking research are: a theoretical framework, simulation studies, and prototype experiments. These contributions provide insight in how foot placement control can improve the robustness, versatility and energy-efficiency of bipedal gait. Regarding robustness, this thesis introduces the theoretical framework of capturability to analyze or synthesize actions that can prevent a fall. Fall avoidance is analyzed by considering N-step capturability: the system's ability to eventually come to a stop without falling by taking N or fewer steps, given its dynamics and actuation limits. Low-dimensional gait models are used to approximate capturability of complex systems. It is shown how foot placement, ankle torques and upper body motions affect the CoM motion and contribute to N-step capturability. N-step capture regions can be projected on the floor: these define where the system can step to remain capturable. The size of these regions can be used as a robustness metric. Regarding versatility, this thesis derives foot placement strategies that enable the system to evolve from the initial state to a desired future state in a minimal number of steps. Simulations on simple gait models demonstrate how these foot placement strategies can be used to change walking speed or walking direction. Regarding energy-efficiency, we learn that simple gait models demonstrate human-like foot placement strategies in response to a stumble when optimizing for either one of the following cost measures for foot placement: peak torque, power, impulse, and torque divided by time. For robotic control, these results indicate that actuator limitations should be taken into account in the execution and planning of foot placement strategies. Regarding robot experiments, we integrate the concepts from the capturability framework into the control of a robot. The low-dimensional gait models are shown to be useful for the robust control of a complex robot. The model takes only the CoM dynamics with respect to the center of pressure (CoP) into account. The application of this model together with force-based control strategies lead to robust robot behavior: upright postural balance is maintained when the robot is pushed and one of the feet is placed on a moving platform. Successful application is also shown for single legged balancing with compensatory stepping to regain balance after a push and (simulated) walking. The main conclusion is that analyzing walking control as a combination of decoupled and low dimensional control tasks allows us to derive simple and useful control heuristics for the control of a complex bipedal robot. We find that the key control task is foot placement, which mostly determines the system's CoM motion by defining possible CoP locations. We can approximate the set of possible foot placement strategies that will not lead to a fall. This set specifies the bounds to which foot placement strategies can be adjusted to achieve more versatile or energy-efficient behavior.

The goal of the research presented in this thesis is to increase the understanding of the human running gait. The understanding of the human running gait is essential for the development of devices, such as prostheses and orthoses, that enable disabled people to run or that enable able people to increase their running performance. Although these devices are currently being developed, there is not much insight yet in the fundamentals of the running gait. This fundamental knowledge is required for improving these devices. One of the big unknowns is how these devices affect the ability of the user to handle disturbances, like sudden pushes or variations in floor height. To gain insight in the fundamentals of the human running gait and the disturbance rejection behavior in particular, the gait synthesis approach is taken. In this approach, the running gait is studied by synthesizing the human running on simulation models and on robots. This allows studying the effects of specific system parameters in a simplified and controlled environment. A number of simulation models and a physical running robot have been developed. The simulation models vary in complexity, from simple simulation models based on the well-known spring loaded inverted pendulum (SLIP) model to simulation models that closely resemble the physical running robot. The simple simulation models are useful to get fundamental insights, due to their simple dynamics. The results of the simple models are validated with the more realistic models and physical running robot. This thesis focuses on the effect of three important system parameters on the disturbance rejection behavior. These three parameters are: the leg stiffness profile, the location of the center-of-mass, and the swing-leg retraction rate. These three parameters were selected, based on our experience with walking robots. The research in this thesis shows that the effects of these parameters are the following. The leg stiffness profile has a significant influence on the disturbance rejection behavior. For a simple running model, we show that nonlinear leg springs can improve the disturbance rejection up to a factor 7 compared to the optimal linear leg spring. The optimal leg stiffness profile for the maximal disturbance rejection behavior is strongly nonlinear. These results show that the generally used linear leg springs are far from optimal in terms of disturbance rejection behavior. The location of the center-of-mass of the torso also has a large influence on the disturbance rejection. The optimal center-of-mass location depends on the type of the expected disturbance, which is above the hip for floor height disturbances and below the hip for push disturbances on the center-of-mass. The commonly used center-of-mass location at the hip is far from optimal. An offset of the center-of-mass location can increase the disturbance rejection up to a factor 10 compared to the center-of-mass at the hip. The swing-leg retraction rate, the speed of the backwards rotation of the front leg prior to touchdown, affects the disturbance rejection rate. We show that this effect is maximal at a mild retraction rate, which is much lower than the retraction rate for ground speed matching. The optimal retraction rate decreases with increasing running velocity. Besides improving disturbance rejection, swing-leg retraction can also reduce energetic losses, impact forces, and the risk of slipping. However, we show that all of the benefits of swing-leg retraction occur at different retraction rates, which indicates that there is an inherent tradeoff to consider when selecting the retraction rate for a robot control system. In addition, the effect of the retraction rate on these benefits is strongly model and/or parameter dependent, making it difficult to make general rules on how to select the retraction rate. Besides the above-mentioned results, this research also revealed the following insights. Firstly, not all results from simple running model studies transfer well to more realistic models and robots. This is especially the case for studies on effects that involve impact dynamics, as impact dynamics greatly depend on the leg morphology. Secondly, the gait sensitivity norm, the disturbance rejection measured introduced by Hobbelen for walking systems, is also suitable for running systems. Finally, the implementation of a spring in parallel with the actuator in the knee joint can greatly reduce the required actuator torque and power. Overall, the results of this thesis show that there are many opportunities to improve the disturbance rejection performance of bipedal running robots. This can be done either by mechanical changes to the robotics system, e.g. implementing a nonlinear leg spring or placing the center-of-mass away from the hip, or by changes to the controller, e.g. implementing swing-leg retraction. The results of this thesis also point out promising directions for the development of better running orthoses and prostheses. Most promising is the implementation of nonlinear springs in exoskeletons, because the results show a large improvement in the disturbance rejection behavior and because nonlinear springs are relatively easy to implement in exoskeletons.

Eight healthy young adult males seated in a rigid chair and restrained by a five point harness belt underwent anterior-posterior random appearing multisine perturbations with a frequency range of 0.3-20 Hz. Six different conditions were tested differentiating in maximum acceleration level ([1;2;4;8] m/s²) and in task (mental arithmetic and blindfolded). The head and neck kinematics were captured by a Qualisys motion capture system and Xsens accelerometers. Muscle activity of the trapezius and sternocleidomastoid muscles was collected by a Delsys EMG system. A two pivot neck model was developed representing the head-neck kinematics separating upper and lower neck kinematics. The kinematics were described with an error margin of 2.5 % of the maximum range of motion of the head. The amount of neck deformation relative to the perturbation is expressed in gain and phase, the linearity is expressed in squared coherence. For the head-neck kinematics a significant (P < 0.05) increase in gain was found for decreasing acceleration levels, indicating non-linearity of the human reflexes and/or the passive neck mechanics. At lower frequencies, the mental arithmetic task resulted in a 9 % decrease of neck deformations (P < 0.01). At lower frequencies, blindfolding resulted in a 16 % increase of neck deformations (P < 0.05). The pivot rotations showed for low acceleration levels similar gain and phase characteristics for both the upper and lower pivot up to 3 Hz. Increasing acceleration levels resulted in a major decrease of relative upper neck deformations (P < 0.05) and an increase of relative lower neck deformations (P < 0.01), suggesting different control strategies for both pivots. For frequencies above approximately 5 Hz an increasing phase lag up to 180º for the upper pivot with respect to the lower pivot is found, indicating C-shaped neck bending for low frequencies and S-shaped neck bending for high frequencies. With the exception of the upper pivot response the squared coherence showed globally values above 0.5 between 1-12 Hz.

In order to understand how the different brain systems work, control theory concepts are used to represent the input - output relationships of the structures involved. Perturbations is a common tool to study and analyze a control system. One example of a brain circuitry is the oculomotor system. Although, visual perturbations have been used to perturb the eye no mechanical perturbations have been applied up to now. Mechanics is the only way to evoke an unexpected movement, which is an essential factor to motor control. In this study, a magnetic actuator is designed to be used to apply torques in the rabbit eye. In vitro experiments were conducted in a prototype, which roughly mimics the movement of the eye in the horizontal plane, to test the function of the actuator. Experiments in the rabbit (in vivo) were also performed. In vitro results showed that the conceptual design is sound and the demanded torque of 17 mN mm was achieved. During preliminary in vivo results, clear eye movements were recorded as a result of the actuator's perturbations. The actuator designed enables a series of experiments in the frame of oculomotor control research.

In this paper, stepping over a zero height obstacle with minimal actuation is studied for a limit cycle walker modeled as a double inverted pendulum. The obstacle position is estimated by stereo vision. Actuation is realized by a constant torque per step on the hip and a push-off collinear to the trailing leg. Stepping over the obstacle must be accomplished with the obstacle position exactly on a predefined position in between the legs with the final state right after push-off being equal to the initial state. Thus, at least two steps must be taken to perform this task, such that the first step is used to make sure the relative position of the obstacle is correct. In the best case scenario, the obstacle is exactly in between the legs during a nominal walk. In that case, actuation does not have to be adjusted with respect to the nominal actuation. In the worst case scenario the obstacle is exactly at a stepping position. In that case, a translation of the step positions is needed. For stepping over in two steps this is not possible; this is only possible for a small range around the best case scenario. For stepping over in three steps this is possible when actuation is applied according to the optimization results presented.

The Dutch Shoulder and Elbow Model (DSEM) is a musculoskeletal model of the shoulder that can be used to predict internal shoulder loading (muscle forces, joint reaction forces, etc.). The DSEM uses an inverse optimisation method to predict muscle forces from net joint moments. In this study two new modes are presented that constrain the inverse optimisation with muscle force boundaries based on muscle dynamics (inverse forward dynamical mode) and boundaries based on EMG-recordings (EMG-assisted mode). The new modes were validated with measurements of two standardised movements (abduction and ante exion) from two subjects. A proof of concept has been given that both new modes work. It was concluded that DSEM predictions can be dominated by morphological differences between the subject and the cadaver on which the DSEM is based. Until better scaling routines are developed the IFDO mode is not very useful. When EMG-constraints are added, muscle and GH-joint reaction forces are predicted to be higher. Adding EMG for one muscle can predict cocontraction in other muscles. By adding EMG-based constraints, the DSEM can account for individual strategies in control strategy for the data that was analysed and is therefore an interesting topic for future research.

Purpose: Executing the sideward skating push-off requires a skater’s full attention and capabilities. Ankle eversion (AE) occurs during the push-off with skaters of low and high skill levels. Controlling AE requires high muscle force. AE adds unnecessary stress thus fatigue is a likely consequence. A purpose for AE in speed skating has not yet been found. Ankle eversion (AE) is considered an unwanted distraction during the execution of the push-off. The first goal is to reduce AE during speed skating. Plantar and dorsal flexion is to be left unhampered. The second goal is to prove that the skating motions can be executed with reduced AE. Method: An orthosis was designed to reduce AE on the right leg only. Skaters (n=10) with low and high skill levels were recorded while skating with normal and reduced AE. Video analyses resulted in relevant angles to quantify skating motions. The tested skaters filled out a questionnaire about skating with reduced AE. Results: On average AE was reduced by 45 to 70% from approximately 13 to 4 degrees with the tested skaters. Skating motions could be executed with reduced AE. The overall rating by the tested skaters for skating with reduced AE was neutral to positive. The orthosis functioned properly but it was considered big and clumsy. Conclusion: There are no negative outcomes from the angle measurements or the questionnaire on skating with reduced AE. That is a very positive situation. An estimation of the required muscle force shows that the amount is reduced significantly when skating with reduced compared to normal AE. As a result a skater saves energy and is not distracted by AE when executing the push-off. Skating with reduced AE might have a positive influence on performance. This has not been measured. A redesign of the orthosis should be stiffer and more compact.

Assessment of the sodium/potassium (Na/K) pump functioning can play an important role in the early diagnosis of several polyneuropathies. However, currently no method exists that allows in vivo testing of the Na/K pump. Therefore, this study developed a test based on repetitive nerve stimulation at single motor units. In order to evaluate the Na/K pump functioning, measurements were taken at various frequencies. Low frequency stimulation was applied to record the recovery cycle of a single motor unit at which no Na/K pump influence is expected. Stimulation trains of physiological frequency perturbed the Na/K pump and led to hyperpolarization of the membrane potential. Hence hyperpolarization leads to a decreased neural excitability, a threshold change can be recorded. Since both stimulation at physiological and low frequency led to a threshold change, a model was introduced to discriminate the influence of the Na/K pump. During the recordings at physiological stimulation frequency a remarkable phenomenon was noticed which was identified as neural bifurcation.

The goal of this study is to quantify the recovery of spinal reflexes at the elbow after neurosurgical intervention in patients with brachial plexus injuries. So far, main focus was on the recovery of muscle force and little on sensory- and reflex system. As reflexes play an important role during normal movement, it is of interest to determine to what extend reflexes have been restored after surgery. Arm admittance (dynamic relationship between displacements in a response to forces) at endpoint level (hand) was estimated using force perturbations in two directions (horizontal) applied by a two-joint robotic manipulator. Three different task instructions were used to provoke different intrinsic and reflexive behavior, being a position task (PT), a relax task (RT) and a force task (FT) where the subject was required to minimize hand displacements, not react to the perturbations and minimize force deviations (being compliant) respectively. Ten patients with brachial plexus lesions participated in this experiment and were suffering from varying degrees of arm dysfunction. All had successful recovery for the biceps (MRC grade 3 and higher) after surgical nerve repair of the n. musculocutaneous. Estimated intrinsic and reflexive parameters were compared to those of a control group (n = 10, age and sex matched). The task instructions had great influence on the admittance, especially between the PT and RT. In all patients, reflexive activity was found corresponding to assumed muscle spindles (velocity- and position feedback) and Golgi tendon organ (force feedback) function. For the PT, the difference in parameters between patients and control subjects was largest. Overall, patients exhibit more intrinsic stiffness at the shoulder and elbow than the control subjects, an indication of co-contraction, and less reflexive feedback at the elbow. There are two possible explanations for this: 1) The intrinsic and reflexive properties did not recover to the combination as before the injury and are not cooperating correctly, and 2) The patients are relying more on intrinsic control than on reflexive control (different control strategy). Whereas the control group uses an energy efficient approach, i.e. less intrinsic and more reflex activity, the patients appear to use a more maximal activation approach resulting in co-contraction. It is possible that the exercises performed during rehabilitation which focus on muscle force do not provoke enough reflexive behavior. More research, e.g. experiments that are designed to disable co-contraction, is needed to verify if the use of co-contraction is learnt or a necessity. Conclusion: reflexes do recover after surgery to severe Brachial plexus injuries, the amount of reflex function is less or less effective than for the control group.

A gravity balancer is a mechanism that compensates the weight of a mass over a range of motion. When no friction is present, this gives an energy efficient mechanism and little effort is required to move an object. Conventional mechanisms have drawbacks due to the use of conventional rigid joints. Compliant joints do not have these disadvantages, can be made from fewer parts and can increase performance compared to rigid body joints. The goal of this paper is to develop a single degree of freedom gravity balancer where all the rigid joints are replaced with compliant joints. To reach this goal a new method has been developed. The method is based on connecting rigid links with compliant joints. With a constant potential energy as objective, the method allows new gravity balancers to be designed. It can be concluded that for the first time a gravity balancers has been constructed where all the rigid joints are replaced with compliant joints. The gravity balancer had a peak moment reduction of 93%. The presented method is extensible and allows others to understand and to further develop gravity balancers with compliant joints for other applications.

Assessing mechanisms of peripheral reflex control is important for understanding movement disorders after suprapsinal nerve lesions like stroke. In the present study, reflex provocation by ramp and hold rotations (R&H) was combined with Transcranial Magnetic Stimulation (TMS). In four subjects, subthreshold single pulses TMS were applied to the primary motor cortex at carefully timed intervals, while short and long latency EMG responses of the m. flexor carpi radialis were elicited by R&H rotations around the wrist joint. TMS was found to inhibit the long latency response with a maximum inhibition when TMS was calculated to arrive at 45ms after stretch onset in all subjects. Excitation was found at 60 ms in all subjects. An involvement of the primary motor cortex in peripheral reflex loop operation was demonstrated. This involvement may be either exictatory or inhibititory on the stretch reflex.

Static balancing is a useful concept to reduce operating effort in mechanisms. A statically balanced system which is designed to counterbalance a mass, is referred to as a gravity equilibrator. The potential energy in a gravity equilibrator is constant, which in most of the times is achieved by mechanical springs. Often helical springs are used, although these springs take a lot of space within the workspace of the mechanism. This paper presents the design of an adjustable gravity equilibrator using torsion bars, which saves space in the working area. Static balancing is achieved with a non-constant transmission (NCT). A new NCT design, and a general method to calculate the design parameters are presented. The stiffness of the torsion bars can be adapted by changing the active length. In this way it is possible to balance different masses with the same system.

An tele-operation, haptic feedback from the remote environment to the human is often limited, which has been shown to negatively influence the performance and required time of tasks. The conventional research focus is on improving the quality of the haptic feedback (transparency), which may have led to significant improvement, but is still imperfect, with many unresolved issues. The present study presents an alternative approach to improve tele-operated tasks: by offering haptic shared control in which both operator and support system apply the required forces at the input (master) device. It is hypothesized that virtual forces from well-designed shared control will improve required time and accuracy, with less control effort, and that these benefits exist for perfect transparency but even more so for imperfect transparency. In an experimental study haptic shared control was designed to aid operators (n=9) with performing a simple bolt-spanner task using a planar (2D, 3DOF) tele-operator setup. Haptic shared control was compared to normal operation for three types of control: the baseline condition of direct control at the master (perfect transparency), teleoperation with a simple PERR controller, and a PERR controller with feedback gains set to zero (no transparency). The experimental results provided evidence for the hypotheses, showing that all tested tele-manipulation tasks benefit from haptic shared control, for all three levels of transparency. Essentially, the presence of haptic shared control allows for a worse transparency without compromising accuracy or required time, and can even improve accuracy and required time during perfect transparency. Subjective results indicated that the shared control was perceived as helpful and beneficial.

The concept of underactuation is increasingly applied to hand prostheses because it minimizes the number of actuators while the fingers are adaptive to the grasped objects. However, current underactuated grasp mechanisms suffer from an intrinsic stability problem which disables them to obtain a pinch grasp, hence they fail to grasp small objects. In an attempt to solve this problem joint locks are applied to an underactuated pulley-tendon driven hand. Joint locks might also have an effect on the power grasp strength. Therefore the effect of joint locks on the pinch grasp performance and power grasp strength of underactuated hand prostheses is investigated. This effect is quantified by two grasp performance metrics: the ability to pinch and the ability to hold. To find the optimal joint lock configurations for pinch grasp performance and power grasp strength, the effect of different joint lock configurations on both performance metrics is calculated using a static grasp model. The results of this model are validated by measurements on a prototype. Both the model and the measurements showed that joint locks do not have a significant effect on the pinch grasp performance when only the range of object sizes for which a pinch grasp equilibrium exists is considered. However, the measurements showed that a pinch grasp is more easily established with joint locks. Additionally, both the model and the measurements showed a significant increase in the power grasp strength.

This three-part paper discusses the analysis and control of legged locomotion in terms of N-step capturability: the ability of a legged system to come to a stop without falling by taking N or fewer steps. We consider this ability to be crucial to legged locomotion and a useful, yet not overly restrictive criterion for stability. Part 1 introduces the theoretical framework for assessing N-step capturability. Formal definitions of N-step capturability and related terms are given, and general disturbance robustness metrics based on capturability are proposed. Part 2 uses the theoretical framework developed in the current part to analyze N-step capturability for three simple gait models. Part 3 describes how the results for the simple models were used to control a complex lower body humanoid robot with two six degree of freedom legs.