1.1 Background and motivation
Since the Berkeley Lower Extremity Exoskeleton (BLEEX) [1] was proposed, the exoskeleton research has become a trend around world. An exoskeleton is designed for strengthening normal human’s behavior. On the other hand, helping weak human to become normal is also a way to strengthen human ability. We focus on lower limb assistance as a mission to allow people to enjoy the freedom of walking. Lower limb problem may happen on human who suffers from serious injury such as car crash or falling. It would also happen on human whose muscle degrades due to aging. The fact that people with degraded muscles are reluctant to exercise makes the lower limb problem even worse. Accordingly, exoskeleton is aimed to modify and to facilitate human locomotion in a strengthened manner.
1.2 Literature review 1.2.1 Skin phantom calibration
The challenge of EEG capturing is that brain signal is weak. How to make reliable measurement is a difficult problem. For resolving this problem, we should make a testing platform, which is repeatable for testing EEG measurement. Using human subject to test EEG recoding has problem that human would not easy to show repeatable brain signal. In addition, using large human subjects to test EEG can be costly. On the other hand, artificial skin phantom can be a reusable test platform. Phantom testing on EEG is promising [2].
In general, artificial skin phantoms can be categorized into two groups, the physical skin phantom and the tissue-engineered skin model. The former is designed to analogize the practical physical condition of a human skin during in vitro experiments, and the latter focuses on the biological relevance of a human skin [3]. Various phantoms of this physical type have been developed for analogizing the mechanical or electrical properties of real human skin. For instance, a gelatin-based model was developed to analogize the mechanical conditions of human skin under reentry shots [4] while a silicone-based model doped with graphite was proposed to analogize the electrical conditions of a human skin [5].
The physical skin phantoms are suitable for our design because we focus on the physical properties of a human skin for the applications of EEG testing and calibration.
The goal of this research is to develop an artificial skin phantom with mechanical and electrical properties of the phantom are close to those of real human skins. The morphology of the phantom is also similar to that of a real human skin. In addition, the artificial skin phantom is prepared for testing EEG device to calibrate noise.
1.2.2 Overview of walking assist device
Many lower limb exoskeletons have been proposed to assist human walking.
Sankai et al. develop hybrid assistive limb-5 (HAL-5) to aid human whose muscle degenerated. HAL-5 have two main control system. One is measuring EMG to know whether human want to walk or not. The other is storing walking patterns for individual user [6]. Long et al. used active disturbance rejection control strategy to track human gait trajectory on a lower limb exoskeleton made of carbon fiber for rehabilitation. They also used an extended-state observer to estimate and then suppress the disturbance by applying a control action [7]. Ollinger et al. developed an admittance control model to
estimate the human-assistive effect on Honda stride management assist (SMA) device.
They also discussed robust stability on their control system [8].
Nagarajan et al. establish control strategy on Honda SMA device, which is based on modify the dynamic response of human limbs. By increasing the mechanical admittance, human lower limbs are more responsive to any muscle torque generated by user [9]. Liu et al. construct variable stiffness actuator, which can modulate output stiffness by changing the effective length of a bending bar. Their results show that the controller can achieve the desired performance in reference tracking [10].
Chen et al. designed a portable knee-ankle-foot robot to help stroke patients. They developed control strategies on gait phase and applied appropriate assistive force at corresponding gait phases [11]. Oh et al. propose various assistive control. Detecting walking motion phases by switching control algorithms, system would pick up appropriate control framework to handle plant [12]. Achili et al. proposed a stable adaptive observer. It can be applied to any other nonlinear system of similar dynamics [13]. Giovacchini et al. presented a lightweight carbon fiber active orthosis, whose weight is 4.2 kg, to assist hip rotation. User could walk with this orthosis without feeling hindered [14]. Asbeck et al. proposed a soft exosuit for portable hip assistance.
Their soft design would not restrict hip ab- and adduction direction or rotation about leg axis [15]. Ouyang et al. developed a power unit for exoskeleton robot, a compact hydraulic power unit powered by an internal combustion engine (CHPU). The CHPU can provide 1.5 kW hydraulic power and 100 W electric power, which can meet the requirement for exoskeleton robots [16]. Selinger et al. designed myoelectric control, which can adapt the timing and magnitude of electrical power generation for an energy harvesting exoskeleton [17]. Hussain et al. proposed an adaptive seamless assist-as-needed (AAN) control, which is for the robotic gait training. They found that
the robotic orthosis is capable of guiding human limbs on reference trajectories [18].
Vouga et al. presented a lower-limb exoskeleton controlled by brain for rhesus macaque.
They demonstrated the feasibility of a brain-controlled lower-limb exoskeleton [19].
Zhang et al. presented a rehabilitation exoskeleton, which can undergo walk training on patient’s individual walking habit. Their exoskeleton can evaluate patient’s rehabilitation status in real time by providing necessary torques on the dyskinetic leg [20]. Jin et al. used adaptive fuzzy sliding mode control on a lower limb exoskeleton.
Their wearer feels more comfortable to move the swing leg [21, 22]. Long et al.
proposed passive mode and active mode, respectively, on a lower limb rehabilitation exoskeleton for unilateral lower limb movement disorders patient. Exoskeleton in the former case would copy healthy gait trajectory. The latter would modify healthy gait trajectory and help to strengthen the unhealthy limb [23]. Zhu et al. presented an unidirectional variable stiffness hydraulic actuator for loading carrying knee exoskeleton. Their results show that the system has good performance on stiffness regulation and joint torque control [24].
We want to develop a compact walking-assistive device on lower limb. Our exoskeleton should let user feel comfortable as receiving the help form the power unit.