Assistive device

Abstract

A powered orthotic device (10) for providing powered assistance to an upper limb of a human user. The device comprises an arm exoskeleton (14), orthoses (28, 30) for mounting the exoskeleton (14) to an upper limb, sensors (36, 38) for measuring torque in the exoskeleton (14), and motorised actuators (42, 44) for controlling movement of the exoskeleton (14). The device (10) also comprises a controller for receiving the sensor output and activating the actuators (42, 44) to control movement of the exoskeleton (14) based on the sensor output.

Description

[0001] ASSISTIVE DEVICE

[0002] TECHNICAL FIELD

[0003] This disclosure relates to powered orthotic devices and methods of controlling such devices. In particular, this disclosure relates to devices and methods for providing powered assistance to an upper limb of a human user.

[0004] BACKGROUND

[0005] Assistive orthotic medical devices exist to support a user in the functional activities of daily living. A user may have a motor impairment in a limb, for example as a result of a medical condition such as a stroke or a brachial plexus injury. Such an impairment may affect an upper limb, meaning that the user’s arm is severely weakened, making several activities very difficult or impossible. Support for the upper limb, to help the user with arm function, may be provided by a passive orthotic device, such as a brace, or a powered orthotic device that uses a power source, such as a battery, to actively support the upper limb and provide powered assistance to limb movements. A powered orthotic device may also be used by people with motor impairment in both upper limbs or by people without motor impairments to enhance their strength or mobility.

[0006] Known powered orthotic devices use electromyography (EMG) to measure the remaining electrical muscle activation signals directly on the skin above a muscle in the impaired arm. This information is used to determine how to provide powered assistance to the elbow or grasp of the user. However, EMG has significant drawbacks for users of powered orthotic devices.

[0007] EMG utilizes electrodes placed on the skin above a muscle. To get sufficient signal quality, the placement of these electrodes must be accurate. The user must place these electrodes every time they are putting the device on, which complicates the process of putting the device on. This process also introduces potential differences in the interaction every time the device is used. Some muscle areas may present particular difficulties for electrode placement by a user - for example, shoulder movements use more muscles than elbow movements, including muscles on the chest and back of the user. These areas may be out of reach and / or sight for the user, and the correct location for electrode placement can be hard to pin-point, even for a skilled operator. EMG is thus not as suitable for controlling powered assistance in shoulder movement.

[0008] Furthermore, EMG electrodes must be placed on exposed skin. This restricts which clothing the user can wear while operating the device. This can be problematic for users with medical conditions that cause blood flow to the extremities to be restricted, resulting in the users being cold.

[0009] WO2016 / 205356A1 discloses a powered orthotic device for a limb that uses EMG sensors to detect activity in a user’s muscles, and applies a torque to a brace system based on the sensor output to assist a user’s motion.

[0010] Some powered orthotic devices are designed to support rehabilitation of an upper limb, rather than to support daily functional activities, and may include a stationary unit (for example installed in a rehabilitation centre) to provide power and communication to a connected orthotic device. For example, the device can be programmed with specific motions for the upper limb to carry out, and sensors can provide feedback about the mobility of the upper limb for tracking rehabilitation exercises. US2009 / 0225620A1 discloses a powered exoskeleton for shoulder rehabilitation in physical therapy.

[0011] There is a need for an improved powered orthotic device to assist users with carrying out functional daily activities.

[0012] SUMMARY

[0013] An aspect of the present disclosure provides a powered orthotic device for providing powered assistance to an upper limb of a human user. The device comprises: an arm exoskeleton comprising an upper arm portion and a forearm portion pivotably connected together by an elbow joint, and a shoulder portion pivotably connected to the upper arm portion by a shoulder joint; one or more orthoses for mounting the exoskeleton to the user’s upper limb; one or more sensors (e.g., torque sensors) at the shoulder joint and / or elbow joint for measuring a torque in the respective shoulder joint and / or elbow joint and providing a sensor output indicating the measured torque; one or more motorised actuators, each configured to produce a torque to control movement of the exoskeleton about the shoulder joint and / or elbow joint to assist the user in carrying out an upper limb movement; and a controller communicatively coupled to the one or more sensors and the one or more motorised actuators. The controller comprises a processor programmed to: receive the sensor output from the one or more sensors; and activate one or more of the one or more motorised actuators to control movement of the exoskeleton about the shoulder joint and / or elbow joint based on the sensor output to assist the user in carrying out the upper limb movement.

[0014] In some embodiments of the above, the processor is programmed to estimate a voluntary joint torque in the user’s shoulder joint and / or elbow joint based on at least the sensor output; determine a control operation for the one or more motorised actuators based on the estimated voluntary joint torque; and activate one or more of the one or more motorised actuators to move the exoskeleton about the shoulder joint and / or elbow joint according to the control operation to assist the user in carrying out the upper limb movement. In some embodiments, the processor is programmed to estimate the voluntary joint torque additionally based on a gravity model defining a torque contribution from at least one of a weight of the user’s upper limb and a weight of the device.

[0015] In some embodiments of any of the above, the one or more sensors comprises a shoulder sensor at the shoulder joint, the shoulder sensor configured to measure a torque in the shoulder joint and provide a shoulder sensor output indicating the torque in the shoulder joint, wherein the one or more motorised actuators comprises a motorised shoulder actuator configured to control the movement of the exoskeleton about the shoulder joint, and wherein the processor is programmed to generate a control signal to move the motorised shoulder actuator based upon at least the shoulder sensor output.

[0016] In some embodiments of any of the above, the one or more sensors comprises an elbow sensor at the elbow joint, the elbow sensor configured to measure a torque in the elbow joint and provide an elbow sensor output indicating the torque in the elbow joint, wherein the one or more motorised actuators comprises a motorised elbow actuator configured to control the movement of the exoskeleton about the elbow joint, wherein the processor is programmed to generate a control signal to move the motorised elbow actuator based upon at least the elbow sensor output.

[0017] In some embodiments of any of the above, the processor is programmed to: estimate a voluntary joint torque (i.e. , a torque exerted by the user) in the user’s shoulder joint based on at least the shoulder sensor output; and determine a control operation for the motorised shoulder actuator based on the estimated voluntary joint torque, wherein the generated control signal is configured to cause the motorised shoulder actuator to operate according to the control operation.

[0018] In some embodiments, if the voluntary joint torque is outside a shouldertorque range, the control operation comprises activating the motorised shoulder actuator to control movement of the upper arm portion about the shoulder joint to assist the user in carrying out the arm movement.

[0019] In some embodiments of any of the above, the processor is programmed to: estimate a voluntary joint torque (i.e., a torque exerted by the user) in the user’s elbow joint based on at least the elbow sensor output; and determine a control operation for the motorised elbow actuator based on the estimated voluntary joint torque, wherein the generated control signal is configured to cause the motorised elbow actuator to operate according to the control operation.

[0020] In some embodiments, if the voluntary joint torque is outside an elbow torque range, the control operation comprises activating the motorised elbow actuator to control movement of the forearm portion about the elbow joint to assist the user in carrying out the arm movement.

[0021] In some embodiments of any of the above, the one or more sensors further comprises an elbow sensor at the elbow joint, the elbow sensor configured to measure a torque in the elbow joint and produce an elbow sensor output indicating the torque in the elbow joint; the one or more motorised actuators further comprises a motorised elbow actuator configured to control the movement of the exoskeleton about the elbow joint; the control signal is a first control signal and the processor is programmed to generate a second control signal to move the motorised elbow actuator based upon at least the elbow sensor output; and the processor is programmed to estimate the voluntary joint torque in the user’s shoulder joint additionally based on at least the elbow sensor output.

[0022] In some embodiments of any of the above, the processor is programmed to estimate the voluntary joint torque additionally based on a gravity model defining a torque contribution from at least one of a weight of the user’s upper limb and a weight of the device.

[0023] In some embodiments, the device further comprises one or more additional sensors coupled to the arm exoskeleton and communicatively coupled to the controller, the one or more additional sensors configured to indicate relative positions of the shoulder joint and elbow joint with respect to each other, or relative angular positions of the upper arm portion and the forearm portion with respect to the gravity vector, wherein the processor is further programmed to receive sensor output from the one or more additional sensors and determine the torque contribution from the at least one of a weight of the user’s upper limb and a weight of the device based on the sensor output from the one or more additional sensors.

[0024] In some embodiments of any of the above, the device further comprises a least one of: a forearm sensor coupled to the forearm portion at a position spaced from the elbow joint, the forearm sensor configured to produce a sensor output indicating a bending moment or transferred shear force in the forearm portion; and an upper arm sensor coupled to the upper arm portion at a position spaced from the elbow joint and the shoulder joint, the upper arm sensor configured to produce a sensor output indicating a bending moment or transferred shear force in the upper arm portion, wherein the processor is programmed to: receive the sensor output from the forearm sensor and / or upper arm sensor; and estimate the voluntary joint torque additionally based on the sensor output from the forearm sensor and / or upper arm sensor.

[0025] In some embodiments, the processor is programmed to use the sensor output the forearm sensor and / or upper arm sensor in combination with the sensor output from the one or more sensors to estimate a distribution of weight supported by the exoskeleton along the upper arm portion and / or forearm portion and estimate a weight of an object held by the user, and the processor is programmed to estimate the voluntary joint torque additionally based on the estimated weight of the object.

[0026] In some embodiments of any of the above, the processor is programmed to determine an output torque or velocity for the shoulder joint and / or elbow joint based on the voluntary joint torque and to control the one or more motorised actuators to produce the output torque or velocity to move the exoskeleton about the respective shoulder joint and / or elbow joint.

[0027] In some embodiments, the output torque or velocity is proportional to the voluntary joint torque.

[0028] In some embodiments, if the magnitude of the voluntary joint torque is below a first predetermined threshold, the output torque or velocity is zero.

[0029] In some embodiments, if the magnitude of the voluntary joint torque is above a second predetermined threshold higher than the first predetermined threshold, the output torque or velocity is zero.

[0030] In some embodiments of any of the above, the device further comprises one or more accelerometers mounted to the device for measuring accelerations of the device, the one or more accelerometers communicatively coupled to the controller, wherein the processor is programmed to receive one or more acceleration measurements from the one or more accelerometers and to estimate the voluntary joint torque additionally based on the one or more acceleration measurements.

[0031] In some embodiments of any of the above, the processor is programmed to switch the one or more motorised actuators to a passive mode, in which movement of the exoskeleton about the shoulder joint and / or elbow joint is permitted without being controlled by the one or more motorised actuators.

[0032] In some embodiments of any of the above, the processor is programmed to switch the one or more motorised actuators to a lock mode such that the one or more motorised actuators prohibit movement of the exoskeleton about the shoulder joint and / or elbow joint.

[0033] In some embodiments of any of the above, the device further comprises a body harness for mounting the device to the user’s torso, wherein the shoulder portion is fastened to the body harness or is integrally formed with the body harness.

[0034] In some embodiments of any of the above, the shoulder joint is a first shoulder joint pivotable about a first shoulder axis and further comprising a second shoulder joint pivotable about a second shoulder axis perpendicular or near perpendicular to the first shoulder axis, wherein the upper arm portion is independently pivotable with respect to the shoulder portion about each of the first shoulder joint and second shoulder joint.

[0035] In some embodiments of any of the above, the device further comprises a gripping portion configured to support the user’s hand, and a gripping actuator configured to open and close the gripping portion to assist the user in opening and closing their hand to grip an object.

[0036] Another aspect of the present disclosure provides method of controlling a powered orthotic device, wherein the device comprises an arm exoskeleton comprising an upper arm portion and a forearm portion pivotably connected together by an elbow joint, and a shoulder portion pivotably connected to the upper arm portion by a shoulder joint, the method comprising: using one or more sensors at the shoulder joint and / or elbow joint, measuring a torque in the respective shoulder joint and / or elbow joint; and activating one or more motorised actuators to control movement of the exoskeleton about the shoulder joint and / or elbow joint based on the measured torque in the respective shoulder joint and / or elbow joint.

[0037] In some embodiments of the above, the method further comprises: estimating a voluntary joint torque (i.e., a torque exerted by the user) in the shoulder joint and / or elbow joint based on the measured torque; determining an output torque or velocity for the one or more motorised actuators based on the voluntary joint torque; activating the one or more motorised actuators to control movement of the exoskeleton about the shoulder joint and / or elbow joint according to the determined output torque or velocity.

[0038] In some embodiments, the output torque or velocity is proportional to the voluntary joint torque.

[0039] In some embodiments, if the magnitude of the voluntary joint torque is below a first predetermined threshold, the output torque or velocity is zero.

[0040] In some embodiments, if the magnitude of the voluntary joint torque is above a second predetermined threshold higher than the first predetermined threshold, the output torque or velocity is zero.

[0041] In some embodiments of any of the above, the method further comprises: using one or more sensors, measuring a bending moment or transferred shear force in the forearm portion and / or upper arm portion; estimating an external weight acting on the device based on the measured bending moment or shear force in the forearm portion and / or upper arm portion; and estimating the voluntary joint torque additionally based on the estimated external weight.

[0042] In some embodiments of any of the above, the method further comprises: defining a gravity model defining a torque contribution from at least one of a weight of the device and external weights acting on the device; and estimating the voluntary joint torque additionally based on the gravity model.

[0043] In some embodiments of any of the above, the method further comprises using one or more sensors to determine relative positions of the shoulder joint and elbow joint with respect to each other, or relative angular positions of the upper arm portion and the forearm portion with respect to the gravity vector, and using the relative positions to define the gravity model.

[0044] In some embodiments of any of the above, the method further comprises: using one or more sensors, measuring an acceleration of the device; and estimating the voluntary joint torque additionally based on the measured acceleration. In some embodiments of any of the above, the method further comprises adjusting control settings of the powered orthotic device via an external communication device in wired or wireless communication with the powered orthotic device.

[0045] In some embodiments, the control settings comprise torque thresholds for controlling activation of the one or more motorised actuators.

[0046] BRIEF DESCRIPTION OF DRAWINGS

[0047] One or more non-limiting examples will now be described, by way of example only, and with reference to the accompanying figures in which:

[0048] Figure 1 shows a powered orthotic device according to an embodiment of the present disclosure, as worn on the torso and upper limb of a user;

[0049] Figure 2 shows the powered orthotic device of Figure 1 without the user’s body;

[0050] Figure 3 shows another view of the powered orthotic device of Figure 1 ;

[0051] Figure 4 shows another view of the powered orthotic device of Figure 1 ;

[0052] Figure 5 shows the same view of the powered orthotic device as Figure 4 and demonstrates how force is modelled in the forearm portion;

[0053] Figure 6 shows graphs demonstrating how the torque contribution of an object held in the user’s hand can be detected and accounted for;

[0054] Figure 7 is a graph showing an example of how actuator output torque or velocity varies with voluntary joint torque;

[0055] Figure 8 is a graph showing another example of how actuator output torque or velocity may vary with voluntary joint torque; and

[0056] Figure 9 shows another view of the powered orthotic device of Figure 1 with an external communication unit.

[0057] DETAILED DESCRIPTION

[0058] In accordance with the present disclosure, Figs. 1 to 5 and 9 show a powered orthotic device 10 for providing powered assistance to an upper limb, i.e., arm and shoulder, of a user to support a user in the functional activities of daily living, such as dressing, cooking and cleaning. The device 10 may be particularly beneficial to users with motor impairments in an upper limb, for example as a result of a medical condition such as a stroke or a brachial plexus injury. However, the device 10 may also be used to augment the upper limb strength and mobility of a user who does not have a perceptible motor impairment.

[0059] The device 10 provides powered assistance to flexion and extension movements of a user’s shoulder and / or elbow. The device 10 may also allow for passive movement in shoulder internal rotation, and may be controllable to lock this movement. The device 10 may also provide support to the user’s hand and fingers for grasping and holding objects (e.g., using a gripping unit 60 as discussed below) and support placement and movement of the user’s wrist joint (e.g., using a wrist unit 62 as discussed below).

[0060] The device comprises a body harness 12 and an arm exoskeleton 14. The body harness 12 allows the exoskeleton 14 to be mounted to the torso of a user, as shown in Fig. 1. The exoskeleton is thus portable and transfers the weight of the arm exoskeleton 14 and the weight of the user’s arm to the torso of the user to make the device 10 easier to wear. The body harness 12 can be custom-made to fit each user. The exoskeleton 14 includes electronics, actuators and sensors for providing powered assistance to the user’s arm, as discussed below. The exoskeleton 14 can be physically attached or mounted to the arm of the user through orthoses on the upper arm and around the wrist and / or hand.

[0061] As shown in Fig. 1 , the exoskeleton 14 comprises a shoulder portion 19, an upper arm portion 20 and a forearm portion 22. One end of the upper arm portion 20 is pivotally coupled to the shoulder portion 19 at a shoulder joint 24. The shoulder portion 19 may be a distinct component in addition to the body harness 12. Alternatively, the shoulder portion 19 may comprise the body harness 12, such that the upper arm portion 20 is coupled directly to the body harness 12 at the shoulder joint 24, or be integrally formed with the body harness 12. The shoulder joint 24 provides relative rotation between the upper arm portion 20 and the shoulder portion 19 about a shoulder flexion axis 16. In some embodiments, the shoulder joint 24 may instead be configured to provide rotation about a shoulder abduction axis or any other axis corresponding to a rotational axis of a human shoulder joint. The other end of the upper arm portion 20 is pivotally coupled to the forearm portion 22 at an elbow joint 26. The elbow joint 26 provides relative rotation between the upper arm portion 20 and the forearm portion 22 about an elbow flexion axis 18. The upper arm portion 20 and shoulder joint 24 correspond to the upper arm and shoulder of the user, and are intended to be aligned therewith when the user is wearing the device 10, as demonstrated in Fig. 1. The forearm portion 22 and elbow joint 26 correspond to the forearm, or forearm, and elbow of the user, and are intended to be approximately aligned therewith when the user is wearing the device 10, as demonstrated in Fig. 1.

[0062] When the exoskeleton 14 is properly tailored and fitted to the user’s arm, the shoulder flexion axis 16 of the shoulder joint 24 of the exoskeleton 14 should be substantially aligned with the flexion axis of the user’s shoulder, and the elbow flexion axis 18 of the elbow joint 26 of the exoskeleton 14 should be substantially aligned with the flexion axis of the user’s elbow. The lengths of the upper arm portion 20 and the forearm portion 22 are known and can be considered to correspond to the lengths of the user’s upper arm and forearm. The distances between the joints 24, 26 and the positions of the orthoses 28, 30, i.e. , the distances between the exoskeleton / user joints and the location of exerted forces, are thus also known and can be used for determining torques in the exoskeleton 14.

[0063] With further reference to Fig. 1 , an additional shoulder pivot joint 34 between the upper arm portion 20 and the shoulder portion 19 or body harness 12 provides relative shoulder internal rotation movement about the shoulder internal rotation pivot axis 32, which is approximately perpendicular to the shoulder flexion axis 16.

[0064] In some embodiments, the body harness 12 is coupled directly to the upper arm portion 20 via the shoulder joint 24 and pivot joint 34. In other embodiments, as shown in Fig. 1 , the shoulder portion 19 is provided between the body harness 12 and upper arm portion 20 such that the body harness 12 is coupled to the shoulder portion 19 via the shoulder pivot joint 34, and the shoulder portion 19 is coupled to the upper arm portion via the shoulder joint 24.

[0065] As shown in Figs. 2 and 3, an upper arm orthosis 28 is mounted to the upper arm portion 20 for engaging the user’s upper arm. Another wrist orthosis 30 is mounted to the distal end of the forearm portion 22 for engaging the user’s forearm, wrist or hand. The orthosis 28, 30 are shown as cuffs, but can take any form that allows engagement with the user’s arm, and may include straps for fastening to the user’s arm. The orthoses 28, 30 can be custom made to fit each user.

[0066] The device 10 allows the user to control powered movements that serve to lift the supported arm, i.e., powered elbow movement and powered shoulder movement. The device 10 operates by recognising the user’s effort to move their arm and providing supplementary power to assist the user in completing the intended movement.

[0067] With reference to Figs. 2 and 3, in this embodiment, the device 10 includes at least three sensors 36, 38, 40 to measure the interaction forces and torques between the user and the exoskeleton 14. The device 10 also includes one or more actuators 42, 44, including motors and a power supply, for producing movement of the exoskeleton 14 about the shoulder and elbow joints 24, 26 to support the user’s arm movement. In some embodiments, the device 10 may have just one or two sensors, or more than three sensors for measuring the interaction forces and torques. Furthermore, the device 10 may have just one of the shoulder actuator 42 and the elbow actuator 44 for producing movement of the exoskeleton 14 about the respective shoulder joint 24 or elbow joint 26.

[0068] A shoulder sensor 36 is provided at the shoulder joint 24 to determine the torque in the shoulder joint 24. For example, the shoulder sensor 36 is directly adjacent to the shoulder joint 24, and / or physically coupled to the shoulder joint 24. Alternatively or additionally, an elbow sensor 38 is provided at the elbow joint 26 to determine the torque in the elbow joint 26. For example, the elbow sensor 38 is directly adjacent to the elbow joint 26, and / or physically coupled to the elbow joint 26 Each of the shoulder sensor 36 and elbow sensor 38 may be any type of sensor capable of providing an output indicating the torque, such as a load cell, for example using one or more strain gauges.

[0069] A shoulder actuator 42 is provided at the shoulder joint 24 for applying torque to the upper arm portion 20 about the shoulder flexion axis 16. Alternatively or additionally, an elbow actuator 44 is provided at the elbow joint 26 for applying torque to the forearm portion 22 about the elbow flexion axis 18.

[0070] In some embodiments, the shoulder sensor 36 is positioned in the shoulder joint 24 and coupled between the shoulder portion 19 and the output of the shoulder actuator 42. When the actuator 42 is not moving, the shoulder joint 24 functions as a stiff joint and the shoulder sensor 36 is thus quasi-rigidly coupled to the upper arm portion 22 so that all torque transmitted through the exoskeleton 14 to the body harness 12 is transmitted through the shoulder sensor 36. This means that forces and torques from the arm of the user and any external force acting on the exoskeleton 14 will be picked up by the shoulder sensor 36.

[0071] In some embodiments, the elbow sensor 38 is positioned in the elbow joint 26 and coupled between the forearm portion 22 and the output of the elbow actuator 44. When the actuator 44 is not moving, the elbow joint 26 functions as a stiff joint and the elbow sensor 38 is thus quasi-rigidly coupled to the upper arm portion 20 so that all torque transmitted through the forearm portion 22 to the upper arm portion 20 is transmitted through the elbow sensor 38. This means that forces and torques from the forearm and hand of the user and any external force acting on the lower part of the arm exoskeleton (below the elbow joint 26) will be picked up by the elbow sensor 38.

[0072] Positioning the shoulder sensor 36 and elbow sensor 38 in this way to measure the joint torques directly provides advantages over deriving or estimating joint torques from interaction forces measured between the user’s arm and the exoskeleton 14 by sensors spaced away from the joints in the forearm portion 22 and upper arm portion 20. Measuring joint torques directly in accordance with the present disclosure enables every interaction force that has a component contributing to a joint torque to be picked up, regardless of the positioning of the user’s arm with respect to the exoskeleton 14 and orthoses 28, 30, or how tightly the arm is fastened to the orthoses 28, 30. Compared to joint torque estimated from interaction forces, directly measured joint torque can more reliably be used as control input to the actuators 42, 44 to provide useful powered assistance to the limb movements.

[0073] As also shown in Figs. 2, 4 and 5, an additional forearm sensor 40 may be provided in the forearm portion 22 at a position offset from the elbow sensor 28 and elbow flexion axis 18. The forearm sensor 40 determines a bending moment or transferred shear force in the forearm portion 22 and may be any type of sensor capable of providing an output indicating the bending moment or transferred force, such as a load cell, for example using one or more strain gauges. For the purposes of measuring the bending moment or transferred shear force in the forearm portion 22, the forearm portion 22 may be divided into least two separate parts coupled together by the sensor 40 positioned between the parts, as demonstrated in Fig. 4.

[0074] The forearm sensor 40 may alternatively be an upper arm sensor coupled to the upper arm portion 20 at a position offset from the elbow sensor 28 and elbow flexion axis 18 and the shoulder sensor 36 and shoulder flexion axis 16. The upper arm sensor determines a bending moment or transferred shear force in the upper arm portion 20. In some embodiments, both a forearm sensor 40 and an upper arm sensor are provided.

[0075] In some embodiments, the device 10 may have just one or two of the three sensors 36, 38, 40. For example, the device 10 may include the shoulder sensor 36 and forearm sensor 40, but no elbow sensor 36, or include the elbow sensor 38 and shoulder sensor 36, but no forearm sensor 40. Alternatively, the device 10 may include one of the shoulder sensor 36 and elbow sensor 38 but not the other two sensors. In embodiments having fewer than three sensors 36, 38, 40, one of the sensors may be configured to produce more than one output signal indicating more than one torque or force in the exoskeleton 14. For example, the sensor functionality of more than one sensor may be provided in a single sensor unit.

[0076] As shown in Fig. 2, the upper arm portion 20 houses a controller 46 for receiving signals from the sensors 36, 38, 40 and controlling the shoulder and elbow actuators 42, 44 based on the sensor signals. The upper arm portion 20 also houses a battery 48 for providing power to the controller 46, sensors 36, 38, 40 and actuators 42, 44. Alternatively, one or both of the controller 46 and battery 48 may be mounted on the body harness 12 or on the forearm portion 22, rather than the upper arm portion 20. Fig. 2 shows the housing of the upper arm portion 20 with an open lid to expose the controller 46, battery 48 and inertial measurement unit (IMU) 47 (discussed below) housed within.

[0077] The controller 46 includes a processor and computer readable memory. The processor is programmed to automatically control the actuators 42, 44 based on the user’s arm movements, as determined by the sensors 36, 38, 40. Input from the sensors 36, 38, 40 allows the controller 46 to recognise the user’s effort to try to move the arm and to control the actuators 42, 44 to provide additional torque to assist the user in completing the desired movement. The shoulder sensor 36 can determine the total torque about the shoulder flexion axis 16 that results from the user’s arm movements and the weights of the user’s arm, the exoskeleton 14, and any object the user is interacting with, e.g., holding in their hand. The elbow sensor 38 can determine the total torque about the elbow flexion axis 18 that results from the user’s arm movements and the weights of the user’s forearm and hand (supported by the exoskeleton 14), the lower part of the exoskeleton 14 beyond the elbow joint 26 and any object in the user’s hand. The forearm sensor 40 can determine the moments resulting from the same sources as the torque determined by the elbow sensor 38. However, the offset placement of the forearm sensor 40 relative to the elbow sensor 38 allows for estimation of the attack point or distribution of the forces, as discussed further below. This information can be used to determine whether the user is holding an object and to estimate the weight of the object. Knowing the weight and / or torque contribution from the object enables the torque from the object to be compensated for in the control of the actuators 42, 44. Without this compensation, the actuators 42, 44 may be controlled to cause movement of the exoskeleton when it is not needed, or provide an unwanted amount of speed in the movement (i.e., move too quickly or too slowly).

[0078] In order to determine the amount of torque with which to supplement the user’s arm movement, the user’s voluntary joint torque (VJT) is estimated for their arm. The VJT is the force exerted by the user’s arm when the user is trying to carry out an arm movement. The VJT includes the magnitude and direction of the user’s applied torque. When the user’s arm is in a relaxed state, the user is not exerting any voluntary force with their arm and so the VJT is zero. If the user moves their arm upwards, this may produce a positive VJT value, and if the user moves their arm downwards, this may produce a negative VJT value. To estimate the VJT, the expected sensor readings from when the user’s arm is in a relaxed state can be determined and stored to allow other sources of torque and moment detected by the sensors 36, 38, 40 to be taken into account.

[0079] For example, when the user exerts a VJT in their elbow when wearing the device 10, the user produces a pushing or pulling force on an orthosis, such as the cuff 30 coupled to the distal end of the forearm portion 22, creating an interaction force between their arm and the orthosis / cuff 30. The VJT in the elbow movement is equal to the interaction force at the cuff multiplied by the distance between the elbow and cuff 30. If there were no other sources of torque, such as gravity (discussed below), this VJT would be equal to the torque measured by the elbow sensor 38.

[0080] The weight of the user’s arm and the weight of the exoskeleton 14 provide additional sources of torque that may affect the sensor readings. To improve the determination of the VJT and thus the accuracy of the required supplementary torque, the weight of the user’s arm and the weight of the exoskeleton 14 can be measured and input to the controller 46. The weight of the exoskeleton 14 includes the weights of the components mounted to the exoskeleton 14, such as the sensors and orthoses. The contribution of the arm and exoskeleton weight to the torque sensor readings can be determined using a gravity model.

[0081] The weight of the forearm portion 22 of the exoskeleton 14 exerts torque on the exoskeleton elbow joint 26. The weight of the forearm portion 22 can be modelled as a point mass with a weight applied at a distance from the elbow joint 26. The torque due to gravity is dependent on the angular position of the forearm portion 22 with respect to the vertical gravity vector, so the angular position of the forearm portion 22 can be determined and input to the gravity model.

[0082] With reference to Fig. 3, one or more position sensors 49 coupled to the controller 46 can be used to determine the angular position of the forearm portion 22. For example, a position sensor 49, such as an encoder, can be used to determine the angular position of the upper arm portion 20 about the shoulder joint 24. A further position sensor 49, such as an encoder, can be used to determine the angular position of the forearm portion 22 relative to the upper arm portion 20. These angular positions can then be combined to determine the angular position of the forearm portion 22, for example with respect to a horizontal plane or the vertical gravity vector. These angular position measurements and calculations may be carried out by the position sensor(s) 49 and controller 46 continuously throughout the user’s arm movements so that the current angular position of the forearm portion 22 is always known to continuously update the gravity model. In the embodiment of Fig. 3, a first position sensor 49 is located at the shoulder joint 24 and a second position sensor 49 is located at the elbow joint 26.

[0083] The torque at the elbow joint 26 due to the weight of the user’s arm can be modelled in a similar way, also using the angular position of the forearm portion 22, which corresponds to the angular position of the user’s forearm. The gravity model thus provides the total torque due to gravity from the sum of the torques due to the weights of the exoskeleton 14 and the user’s arm. The torque contributions from the exoskeleton weight and weight of the arm of the user can also be calculated for the shoulder.

[0084] In addition to the known weight of the exoskeleton 14 and the known weight of the arm of the user, other sources of joint torques in the arm exoskeleton 14 that may be taken into account in the gravity model include: an applied shoulder torque from the user, an applied elbow torque from the user, and any interaction between the user and the environment, e.g., the weight of an object the user is holding in their hand. To perform activities of daily living with the device 10, it is beneficial for the user to be able to control the movements of the device 10 reliably even when holding an object in their hand. To control the device 10 even when holding an object, the controller 46 can use the sensors and stored data about the user’s arm and the device (e.g., weights, dimensions) to determine whether or not the user is holding an object, and estimate the weight of this object.

[0085] To determine if the user is holding an object in their hand, and thus determine the torque contribution from the object, the forearm sensor 40, and / or the upper arm sensor, may be provided to identify a bending moment in the exoskeleton 14. With reference to Figs. 2 and 4, the forearm sensor 40 is located at position between the elbow sensor 38 and the cuff 30 at the distal end of the forearm portion 22.

[0086] The forearm sensor 40 measures the torque transferred through the forearm portion 22 (the bending moment). The forearm sensor 40 is placed at a distance from the elbow sensor 38 so that the third torque measurement point along the forearm portion 22 is at a distance from the elbow measurement point at the elbow joint 26. The forearm sensor 40 is located at a fixed position between the cuff 30 and the elbow sensor 38. The distances between the elbow sensor 38 and the wrist 30 (e.g., a central point along the length of the cuff 30), and between the forearm sensor 40 and the elbow sensor 38 are constant and can be measured and stored within the memory of the controller 46. The distance between the elbow sensor 38 and an object in the user’s hand can be estimated based on the distance between the user’s hand and elbow sensor 38, which can be measured when the user is wearing the device 10 and input to the controller 46.

[0087] With reference to Fig. 5, the output of the elbow sensor 38 indicates the torque in the elbow joint 26 and the output of the forearm sensor 40 indicates the torque in the forearm portion 22 at a distance r from the elbow joint 26. These two torque values, and the known distance r between the measurements, give enough information to estimate the additional torque due to the weight of an object in the user’s hand. To determine whether the user is holding an object or not, the torque about the elbow joint 26 measured by the elbow sensor 38 is modelled as a simple point force F multiplied by a distance d from the elbow joint 26. The magnitude of the torque measured by each sensor 38, 40 is determined, and the distance r between the sensors 38, 40 is known. Based on these values, the value of the modelled point force F and the corresponding value of the distance d can be calculated.

[0088] The distance d can be calculated as: where M38 is the torque / moment measured by the elbow sensor 38 and M40 is the torque / moment measured by the forearm sensor 40. The point force F can further be calculated as:

[0089] When the user is applying a voluntary torque in the shoulder joint 24 or elbow joint 26, this distance d from the elbow joint 26 to where the interaction force is applied is known and within a range.

[0090] When the user is grasping an object, and thus supporting its weight, the point force F increases by AF and the distance d increases by Ad and approaches the distance of the hand of the user from the elbow joint 26. The change in the value of d, Ad, is constantly monitored and when a significant change, for example a change above a pre-set threshold, occurs an object is assumed to be grasped. The change in the modelled force AF is used to estimate the weight of the object. Further data can be utilized to update and improve the estimate of object weight.

[0091] The approximate placement of an object held by the user relative to the exoskeleton is known, i.e. , in the user’s hand (the user’s limb dimensions can be measured and stored in advance). Knowing the approximate object weight and location, the torque contribution from the object to the elbow joint 26 and shoulder joint 24, i.e., measured by the shoulder and elbow sensors 36, 38, can be calculated. This contribution can be removed to improve the estimate of VJT.

[0092] Fig. 6 shows graphs demonstrating how the outputs of the elbow sensor 38 and forearm sensor 40 can be used to adjust the VJT according to the gravity model. Each graph shows time along the x-axis, with t1 corresponding to when a user voluntarily applies a joint torque in their arm, for example in an elbow extension movement, and t2 corresponding to the user picking up and then releasing an object. The combination of the forearm sensor 40 with the shoulder torque sensor 36 and the elbow torque sensor 38 thus enables the additional torque resulting from the user holding an object to be distinguished from an applied shoulder or elbow torque from the user. Graph A demonstrates the change in the moment / torque M38, M40 measured by each of the elbow sensor 38 and forearm sensor 40 at each of t1 and t2. Graph B shows how the distance d between the elbow sensor 38 and the location of the modelled point force F stays below a predetermined threshold distance d1 while the user is just moving their arm at t1 , and increases above the threshold d1 when the user picks up an object at t2. The increase in d beyond the threshold is used to determine that the user is holding an object.

[0093] Graph C demonstrates the change in the magnitude of the modelled point force F as the user moves their arm at t1 and picks up and releases an object at t2. When the user picks up an object, which is determined by the distance d increasing (graph B), the change in the force AF can be used to estimate the weight of the object in the user’s hand. This allows the weight of the object to be compensated for in the VJT, as demonstrated in graph D.

[0094] Graph D shows the change in the determined VJT as the user moves their arm at t1 and picks up and releases an object at t2. A predetermined threshold VJT1 is shown which may be used to determine when to operate the actuators 42, 44 to supplement the user’s arm movement, as discussed further below with reference to Figs. 7 and 8. For instance, the threshold VJT 1 may correspond to the lower threshold 82 in the graphs of Figs. 7 and 8. As can be seen at t1 , when the user is exerting a torque in their arm, this is recognised in the VJT shown by line 41 , and the actuators 42, 44 can be activated when the VJT is above the threshold VJT1 to assist the user’s movement. However, at t2, because the increased distance Ad has been recognised and it has been determined that an object has been picked up, the additional measured torque shown by the dotted line 43 is omitted from the VJT 41 so that the VJT 41 does not go above the threshold VJT1 , which prevents the actuators 42, 44 from being activated unnecessarily.

[0095] Another source of torque that may contribute to the sensor readings is inertial forces stemming from body movements of the user. The inertial forces may be accounted for by comparing patterns in the sensor signals with known patterns which represent normal human movement (e.g., frequency and duration of movements). For example, by carrying out testing, readings from the sensors 36, 38, 40 when a user is carrying out certain body movements may be recorded and used to generate a model or library of human movement patterns. Subsequent sensor readings can then be filtered to recognise known patterns and choose whether or not to take the resulting torques / moments into account when estimating the VJT.

[0096] An inertial measurement unit (IMU) 47 comprising one or more sensors for measuring accelerations in three dimensions (x, y, and z) can be coupled to the exoskeleton 14, or the body harness 12 and be communicatively coupled with the controller 46 to measure and communicate the acceleration of the device 10. The acceleration can be utilized by the controller 46 to recognize inertial forces to improve the estimate of VJT. The acceleration measurements from the IMU 47 can additionally or alternatively be utilized to estimate the direction of the gravity vector to improve the gravity model and correspondingly the estimation of VJT. Any accelerometer or sensor capable of measuring accelerations can be utilized. As shown in Fig. 2, the IMU 47 may be housed within the upper arm portion 20 along with the controller 46 and battery 48. By determining the amount of torque that does not originate from a voluntary arm movement of the user, for example by accounting for known inertial movement patterns and / or using the gravity model, a more accurate VJT contribution can be estimated, e.g., by subtracting the determined additional sources of torque from the measured torque. Individual VJTs for the shoulder joint 24 and the elbow joint 26 can be used to determine the powered assistance to be provided to each joint 24, 26 by the motorised actuators 42, 44. The VJT can be continually and repeatedly estimated so that changes in the user’s intent can be constantly recognised and the most useful assistance can be provided.

[0097] Based on the continuously estimated VJT signal, an intent recognition algorithm can be used to estimate whether the user wants to move, and which level of assistance is needed. The movement control output from the actuators 42, 44 can be calculated based on the application of the intent recognition algorithm on the estimated VJT. For example, if the VJT for the shoulder joint 24 or elbow joint 26 is estimated to be zero, then it may be determined that the user’s movement intent is to keep that joint in a resting position. If the VJT is estimated to be positive, then it may be determined that the user’s movement intent is to make an upward movement of the joint. If the VJT is estimated to be negative, then it may be determined that the user’s movement intent is to make a downward movement of the joint.

[0098] The intent recognition algorithm utilizes the estimated VJT signal in its raw form, and also after applying a high-pass filter. The high-pass filtered signal shows abrupt changes in the value which can indicate changes of intent. Additionally, the estimated VJT and high- pass filtered VJT are compared to one or more predetermined torque thresholds to recognise intent. The threshold(s), which define ranges, enables small values of VJT to be ignored. For example, a minimum magnitude of VJT, i.e. , a minimum torque value above or below 0, may be required to effect operation of a motorised actuator.

[0099] The threshold(s) for the VJT magnitude may be predetermined at least in that it is determined for a comparison of the most recent estimate of the VJT. The predetermined threshold(s) may be a fixed value that is pre-programmed into the controller 46. Alternatively, or additionally, the threshold(s) may be updated in real-time by the controller 46, for example depending on the previously estimated values of the VJT. The updated threshold(s) is thus still “predetermined” by the processor for making the comparison with the estimated VJT.

[0100] The intent recognition algorithm may use knowledge about how humans move (e.g., frequency and duration of movements) to further increase accuracy. This knowledge can be obtained through testing procedures with human users. Movement modes of a user include: initiation, moving, stopping and resting. For example, the VJT and high pass filtered VJT may be expected to be around zero when the user is resting. Each of the movement modes produces patterns in the VJT and high pass filtered VJT that can be observed and used to recognize a change of intent. The estimated VJT and high-pass filtered VJT can be compared to values which would be expected (e.g., established through testing) within these different movement patterns.

[0101] Based on the estimated direction and magnitude of the VJT, the motorised actuators 42, 44 can be operated to supplement the VJT with additional torque to assist the user in carrying out their intended arm movement. The operation of the actuators 42, 44 is dependent on the VJT, i.e., dependent on the intentions / actions of the user. For example, for each joint 24, 26, the output torque or velocity from the actuator 42, 44 may be proportional to the estimated VJT. This allows the device 10 to be adaptive in assisting the user in carrying out normal activities as they go about their daily life while wearing the device 10. The movement of the joints 24, 26, as controlled by the actuators 42, 44 may be produced based on a time series ofVJTs. The current estimated VJT, previous values of estimated VJT, and the change in estimated VJT are used to derive how the actuator 42, 44 should move.

[0102] The controller 46 is programmed to operate one or both of the actuators 42, 44 to move the upper arm portion 20 and / or forearm portion 22 in the direction in which it is determined that the user is exerting force with their arm. The controller 46 may also be programmed to control the actuators 42, 44 to hold the exoskeleton 14 in a resting position, for example by locking the actuators in a fixed position (e.g., with a brake mechanism) to prevent movement of the exoskeleton 14 about the joints 24, 26, or by producing torque to oppose the force of gravity, for example when the user is exerting a VJT in an attempt to hold their arm up and away from their body.

[0103] The amount of torque or velocity that each actuator 42, 44 is controlled to supply to the exoskeleton movements may be proportional to the magnitude of the estimated VJT. For example, if the estimated VJT is relatively low, then the actuator 42, 44 may provide a relatively small amount of torque or velocity. Conversely, if the estimated VJT is relatively high, then the actuator 42, 44 may correspondingly provide a relatively high amount of torque.

[0104] The graph shown in Fig. 7 provides an example of how the output torque or velocity (y-axis) from one or both of the actuators 42, 44 may be controlled depending on the estimated VJT (x-axis) for one or both joints 24, 26. The graph shows upper and lower thresholds 80, 82 that set requirements on the VJT magnitude for obtaining actuator output. The VJT, and consequent actuator output, may be positive or negative, depending on the determined direction of the user’s movement. As shown in Fig. 7, the thresholds are set at equal VJT magnitudes in both directions. However, the positive and negative VJT thresholds 80, 82 do not have to be set at the same magnitudes and can be different to each other.

[0105] For estimated VJTs below the lower VJT threshold 82 and above the upper threshold 80, the actuator output torque / velocity is zero. Between the upper and lower VJT thresholds 80, 82, the actuators 42, 44 are controlled to produce an output torque. The output torque / velocity is set to a constant value for all values of the estimated VJT between the thresholds 80, 82. The lower threshold 82 helps to account for noise in the estimated VJT when the user’s arm is substantially at rest. The lower threshold 82 can be calibrated for different users, for example it may be set to a higher value for users who experience uncontrolled limb movements when at rest to help avoid activating the actuators when the user is not intending to make an arm movement. The upper threshold 80 enables unexpectedly large VJTs to be discounted. For example, if the estimated VJT is too large, there may have been an external or involuntary force which was not accounted for and which falsifies the VJT estimate. The upper and / or lower thresholds 80, 82 can be pre-programmed into the controller 46 and may remain fixed at constant values, or may be varied in real-time, for example depending on previous values of the estimated VJT or the user’s preceding movement pattern. The VJT thresholds 80, 82 can vary between different users and can be calibrated to a specific user’s requirements, for example depending on their strength, mobility, and / or medical conditions. The graph in Fig. 8 provides another example of how the output torque or velocity (y- axis) from one or both of the actuators 42, 44 may be controlled depending on the estimated VJT (x-axis) for one or both joints 24, 26. As well as the upper and lower VJT thresholds 80, 82 also shown in the graph of Fig. 7, the graph of Fig. 8 demonstrates a threshold 84 for the maximum actuator output torque / velocity. This maximum output threshold 84 puts a limit on the torque / velocity that the actuators 42, 44 will output, irrespective of the estimated VJT. This helps to ensure safe and reliable operation of the device 10 and helps to protect the user from injury and the device 10 from damage if the output torque or velocity were to be too high. The maximum output threshold 84 may be set at the same magnitude of torque or velocity for both directions of movement, i.e., for both positive and negative values of output torque / velocity. Alternatively, the magnitude of the output threshold 84 may differ between directions of movement. The maximum output threshold 84 can be pre-programmed into the controller 46 and may remain fixed at a constant value, or may be varied in real-time, for example depending on previous values of the estimated VJT or the user’s preceding movement pattern. The maximum output threshold 84 can vary between different users and can be calibrated to a specific user’s requirements, for example depending on their strength, mobility, and / or medical conditions.

[0106] The graph of Fig. 8 also shows that the output torque / velocity does not have to be constant between the upper and lower thresholds 80, 82 and may vary in a linearly proportional manner with the estimated VJT. As shown in Fig. 8, the output torque / velocity increases with increasing estimated VJT until the output is limited by the maximum output threshold 84. The output then continues to be produced at the constant maximum value as the VJT continues to increase until the upper VJT threshold 80. The output may also vary with the VJT in a non-linear manner.

[0107] Alternatively, or in addition to, programming the processor with one or more set intent recognition algorithms, the processor may use machine learning to estimate the voluntary joint torque and / or control operation of the motorised actuators. For example, to improve estimation of the voluntary joint torque, torque sensor data with different disturbances (e.g., spastic user with involuntary movements, or a user walking) combined with machine learning can be utilized to develop algorithms that can distinguish voluntary movements from such disturbances.

[0108] Machine learning approaches could be used to tune the intent recognition algorithm, for example in classifying typical human movement patterns. The intent recognition algorithm may use knowledge of human movement patterns to define probabilities of being in movement states or transitioning from one movement state to another. These probabilities may be further elaborated through machine learning.

[0109] In some embodiments, the shoulder actuator 42 and / or the elbow actuator 44 can be operated in different modes depending on the amount and / or type of assistance required by the user. In a power augmentation mode, i.e., normal operation of the device 10, the shoulder actuator 42 and / or the elbow actuator 44 are controlled by the controller 46 in response to the estimated VJT to apply torque to move the respective upper arm portion 20 and / or forearm portion 22 about the respective shoulder joint 24 and / or elbow joint 26.

[0110] When the user’s arm is at rest and no active movement assistance is required from the actuators 42, 44, the controller 46 may put the shoulder actuator 42 and / or the elbow actuator 44 in a lock mode, whereby the shoulder actuator 42 and / or the elbow actuator 44 are locked in a fixed position (e.g., with a brake mechanism) to prevent movement of the exoskeleton 14 about the respective joints 24, 26, as discussed above when the user’s arm is at rest. This may be beneficial for users whose limbs undergo involuntary movements due to muscle spasms. The lock mode may be activated automatically by the controller 46 when a resting movement mode is detected for the user’s arm. This means that the actuators 42, 44 do not have to continuously provide a torque to hold the arm at rest in an elevated position. Alternatively or additionally, the lock mode may be activated manually by the user, for example via an external communication unit 70, discussed below. Manual activation of the lock mode may be particularly useful for a user who is unable to control their arm movements, and thus is unable to signal to the controller 46 using their VJT that a resting mode is desired. The lock mode also allows the user to put weight on their arm without the arm moving, for example for leaning on the arm.

[0111] The lock mode of each actuator 42, 44 may be controlled independently. For example, the user may wish to have the elbow actuator 44 in the lock mode but the shoulder actuator 42 in the power augmentation mode in order to lock their elbow movement but still receive powered assistance to their shoulder movement, or vice versa, depending on their individual mobility requirements.

[0112] Alternatively or additionally, the shoulder actuator 42 and / or the elbow actuator 44 may also be operated in a passive mode in which they allow but do not control movement of the exoskeleton 14 about the respective joints 24, 26. The passive mode may be activated manually by the user, for example via an external communication unit 70, discussed below. In the passive mode, the actuators 42, 44 are disabled so that the user can freely move their arm, or their arm can be moved by external forces, without any powered assistance. When the actuators 42, 44 are in the passive mode, the sensors 36, 38, 40, 49 may still be operable.

[0113] The passive mode of each actuator 42, 44 may be controlled independently, for example to enable powered assistance to be selectively applied to each joint 24, 26. The user may wish to have the elbow actuator 44 in the passive mode but the shoulder actuator 42 in the power augmentation mode in order to have free control over their elbow movement but still receive powered assistance to their shoulder movement, or vice versa, depending on their individual mobility requirements.

[0114] In some embodiments, the actuators 42, 44 may be independently controlled to be in one of the lock mode and passive mode. For example, the elbow actuator 44 may be placed in the lock mode, and the shoulder actuator 42 may be placed in the passive mode, or vice versa.

[0115] With particular reference to Fig. 4, in addition to Figs. 1 to 3, in some embodiments, the device 10 may comprise a gripping unit 60 for providing support in opening and closing the user’s hand to grasp and hold objects. The gripping unit 60 comprises a hand portion 50, a finger portion 52 and a thumb portion 56. One end of the hand portion 50 is coupled to the distal end of the forearm portion 22. The other end of the hand portion 50 is pivotally coupled to the finger portion 52 at a finger joint 54. The finger joint 54 provides relative rotation between the hand portion 50 and the finger portion 52. The hand portion 50 may be fitted to the hand of a user such that the finger joint 54 allows the finger portion 52 to approximately align with the user’s fingers during a closing motion of the user’s fingers. The finger portion 52 may be mounted to the fingers of the user such that the relative rotation of the finger portion 52 and the hand portion 50 about the finger joint 54 engages movement of the fingers of the user relative to the hand of the user.

[0116] The thumb portion 56 is coupled to the hand portion 50 and may be oriented such that it provides an opposing surface to a surface of the pivoting finger portion 52 such that the finger portion 52 and thumb portion 56 can meet in a claw-like or pincer-like fashion to provide a grip. The thumb portion 56 may be fitted to the hand of the user such that it supports the thumb of the user in a functional position for gripping. The opposing surfaces of the thumb portion 56 and pivotal finger portion 52 together provide a grasp that can be utilized to grab and hold on to objects. A finger actuator 58 may be provided at the finger joint 54 for applying torque to the finger portion 52 about the finger joint 54 to assist the user in closing or opening their fingers to form a grip.

[0117] The finger actuator 58 can be controlled by the controller 46 to apply a torque to open or close the grip, and to keep the grip open or closed. The controller 46 can obtain the current gripping state (closed or open) of the gripping unit 60, for example by keeping track of the commands sent to the gripping unit 60 and / or by communicating with a gripping sensor capable of determining the state of the gripping unit 60. A gripping sensor may comprise a position sensor capable of determining the relative proximity of the finger portion 52 and thumb portion 56. In some embodiments, the controller 46 can utilize information about the gripping state of the gripping unit 60 to further improve the estimation of whether the user is gripping an object. For example, if the gripping unit 60 is in a closed grip state, the probability that the user is gripping an object may be higher.

[0118] With particular reference to Figs. 1 and 4, in some embodiments, the device 10 may additionally or alternatively include a motorised wrist unit 62 configured to support and / or position the wrist joint of the user in a functional position to perform activities of daily living.

[0119] With reference to Fig. 9, in some embodiments, an external communication unit or device 70 communicatively coupled to the controller 46 can be used to adjust settings on the device 10. The external communication unit 70 communicatively may be coupled to the controller 46 via a wired connection 72 and / or via a wireless connection 74. For example, the external communication unit 70 may be configured to change thresholds on the device 10, change calibration parameters related to the gravity model, and / or change how the output is mapped to VJT. The external communication unit 70 may be in the form of a mobile application, for example on a smartphone or tablet, in wireless communication with the controller 46, a computer program on an external device in wired communication with the controller 46, or any other known communication unit.

Claims

CLAIMS1 . A powered orthotic device (10) for providing powered assistance to an upper limb of a human user, the device comprising: an arm exoskeleton (14) comprising an upper arm portion (20) and a forearm portion (22) pivotably connected together by an elbow joint (26), and a shoulder portion (19) pivotably connected to the upper arm portion (20) by a shoulder joint (24); one or more orthoses (28, 30) for mounting the exoskeleton (14) to the user’s upper limb; one or more sensors (36, 38) at the shoulder joint (24) and / or elbow joint (26) for measuring a torque in the respective shoulder joint (24) and / or elbow joint (26) and providing a sensor output indicating the measured torque; one or more motorised actuators (42, 44), each configured to produce a torque to control movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) to assist the user in carrying out an upper limb movement; and a controller (46) communicatively coupled to the one or more sensors (26, 38) and the one or more motorised actuators (42, 44), the controller (46) comprising a processor programmed to: receive the sensor output from the one or more sensors (36, 38); and activate one or more of the one or more motorised actuators (42, 44) to control movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) based on the sensor output to assist the user in carrying out the upper limb movement.

2. The device of claim 1 , wherein the processor is programmed to: estimate a voluntary joint torque in the user’s shoulder joint (24) and / or elbow joint (26) based on at least the sensor output; determine a control operation for the one or more motorised actuators (42, 44) based on the estimated voluntary joint torque; and activate one or more of the one or more motorised actuators (42, 44) to move the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) according to the control operation to assist the user in carrying out the upper limb movement.

3. The device of claim 2, wherein the processor is programmed to estimate the voluntary joint torque additionally based on a gravity model defining a torque contribution from at least one of a weight of the user’s upper limb and a weight of the device.

4. The device of claim 1 or 2, wherein the one or more sensors (36, 38) comprises a shoulder sensor (36) at the shoulder joint (24), the shoulder sensor (36) configured to measure a torque in the shoulder joint (24) and provide a shoulder sensor output indicating the torque in the shoulder joint (24), wherein the one or more motorised actuators (42, 44) comprises a motorised shoulder actuator (42) configured to control the movement of the exoskeleton (14) about the shoulder joint (24), and wherein the processor is programmed to generate a control signal to move the motorised shoulder actuator (42) based upon at least the shoulder sensor output.

5. The device of claim 1 , 2 or 4, wherein the one or more sensors (36, 38) comprises an elbow sensor (38) at the elbow joint (26), the elbow sensor (38) configured to measure a torque in the elbow joint (26) and provide an elbow sensor output indicating the torque in the elbow joint (26), wherein the one or more motorised actuators (42, 44) comprises a motorised elbow actuator (44) configured to control the movement of the exoskeleton (14) about the elbow joint (26), wherein the processor is programmed to generate a control signal to move the motorised elbow actuator (44) based upon at least the elbow sensor output.

6. The device of claim 4, where the processor is programmed to: estimate a voluntary joint torque in the user’s shoulder joint (24) based on at least the shoulder sensor output; and determine a control operation for the motorised shoulder actuator (42) based on the estimated voluntary joint torque, wherein the generated control signal is configured to cause the motorised shoulder actuator (42) to operate according to the control operation.

7. The device of claim 6, wherein, if the voluntary joint torque is outside a shoulder torque range, the control operation comprises activating the motorised shoulder actuator (44) to control movement of the upper arm portion (20) about the shoulder joint (24) to assist the user in carrying out the arm movement.

8. The device of claim 5, where the processor is programmed to: estimate a voluntary joint torque in the user’s elbow joint (26) based on at least the elbow sensor output; and determine a control operation for the motorised elbow actuator (44) based on the estimated voluntary joint torque, wherein the generated control signal is configured to cause the motorised elbow actuator (44) to operate according to the control operation.

9. The device of claim 8, wherein, if the voluntary joint torque is outside an elbow torque range, the control operation comprises activating the motorised elbow actuator (44) to control movement of the forearm portion (22) about the elbow joint (26) to assist the user in carrying out the arm movement.

10. The device of claim 6 or 7, wherein: the one or more sensors (36, 38) further comprises an elbow sensor (38) at the elbow joint (26), the elbow sensor (38) configured to measure a torque in the elbow joint (26) and produce an elbow sensor output indicating the torque in the elbow joint (26); the one or more motorised actuators (42, 44) further comprises a motorised elbow actuator (44) configured to control the movement of the exoskeleton (14) about the elbow jointthe control signal is a first control signal and the processor is programmed to generate a second control signal to move the motorised elbow actuator (44) based upon at least the elbow sensor output; and the processor is programmed to estimate the voluntary joint torque in the user’s shoulder joint (24) additionally based on at least the elbow sensor output.11 . The device of any of claims 6 to 10, wherein the processor is programmed to estimate the voluntary joint torque additionally based on a gravity model defining a torque contribution from at least one of a weight of the user’s upper limb and a weight of the device.

12. The device of claim 3 or 11 , further comprising one or more additional sensors (49) coupled to the arm exoskeleton (14) and communicatively coupled to the controller (46), the one or more additional sensors (49) configured to indicate relative positions of the shoulder joint (24) and elbow joint (26) with respect to each other, and / or relative angular positions of the upper arm portion (20) and the forearm portion (22) with respect to the gravity vector, wherein the processor is further programmed to receive sensor output from the one or more additional sensors (49) and determine the torque contribution from the at least one of a weight of the user’s upper limb and a weight of the device based on the sensor output from the one or more additional sensors (49).

13. The device of any of claims 2, 3 and 6 to 12, further comprising a least one of: a forearm sensor (40) coupled to the forearm portion (22) at a position spaced from the elbow joint (26), the forearm sensor (40) configured to produce a sensor output indicating a bending moment or transferred shear force in the forearm portion (22); and an upper arm sensor coupled to the upper arm portion (20) at a position spaced from the elbow joint (26) and the shoulder joint (24), the upper arm sensor configured to produce a sensor output indicating a bending moment or transferred shear force in the upper arm portion (20), wherein the processor is programmed to: receive the sensor output from the forearm sensor (40) and / or upper arm sensor; and estimate the voluntary joint torque additionally based on the sensor output from the forearm sensor (40) and / or upper arm sensor.

14. The device of claim 13, wherein the processor is programmed to use the sensor output the forearm sensor (40) and / or upper arm sensor in combination with the sensor output from the one or more sensors (36, 38) to estimate a distribution of weight supported by the exoskeleton (14) along the upper arm portion (20) and / or forearm portion (22) and estimate a weight of an object held by the user, and the processor is programmed to estimate the voluntary joint torque additionally based on the estimated weight of the object.

15. The device of any of claims 2, 3 and 6 to 14, wherein the processor is programmed to determine an output torque or velocity for the shoulder joint (24) and / or elbow joint (26) based on the voluntary joint torque and to control the one or more motorised actuators (42,44) to produce the output torque or velocity to move the exoskeleton (14) about the respective shoulder joint (24) and / or elbow joint (26).

16. The device of claim 15, wherein the output torque or velocity is proportional to the voluntary joint torque.

17. The device of claim 15 or 16, wherein, if the magnitude of the voluntary joint torque is below a first predetermined threshold, the output torque or velocity is zero.

18. The device of claim 17, wherein, if the magnitude ofthe voluntary joint torque is above a second predetermined threshold higher than the first predetermined threshold, the output torque or velocity is zero.

19. The device of any of claims 2, 3 and 6 to 18, further comprising one or more accelerometers (47) mounted to the device (10) for measuring accelerations of the device (10), the one or more accelerometers (47) communicatively coupled to the controller (46), wherein the processor is programmed to receive one or more acceleration measurements from the one or more accelerometers (47) and to estimate the voluntary joint torque additionally based on the one or more acceleration measurements.

20. The device of any preceding claim, wherein the processor is programmed to switch the one or more motorised actuators (42, 44) to a passive mode, in which movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) is permitted without being controlled by the one or more motorised actuators (42, 44).21 . The device of any preceding claim, wherein the processor is programmed to switch the one or more motorised actuators (42, 44) to a lock mode such that the one or more motorised actuators (42, 44) prohibit movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26).

22. The device of any preceding claim, further comprising a body harness (12) for mounting the device to the user’s torso, wherein the shoulder portion (19) is fastened to the body harness (12) or is integrally formed with the body harness (12).

23. The device of any preceding claim, wherein the shoulder joint (24) is a first shoulder joint (24) pivotable about a first shoulder axis (16) and further comprising a second shoulder joint (34) pivotable about a second shoulder axis (32) perpendicular or near perpendicular to the first shoulder axis (16), wherein the upper arm portion (20) is independently pivotable with respect to the shoulder portion (19) about each of the first shoulder joint (24) and second shoulder joint (34).

24. The device of any preceding claim, further comprising a gripping portion (60) configured to support the user’s hand, and a gripping actuator (58) configured to open andclose the gripping portion (60) to assist the user in opening and closing their hand to grip an object.

25. A method of controlling a powered orthotic device (10) for an upper limb of a human user, wherein the device (10) comprises an arm exoskeleton (14) comprising: an upper arm portion (20) and a forearm portion (22) pivotably connected together by an elbow joint (26); and a shoulder portion (19) pivotably connected to the upper arm portion (20) by a shoulder joint (24), the method comprising: using one or more sensors (36, 38) at the shoulder joint (24) and / or elbow joint (26), measuring a torque in the respective shoulder joint (24) and / or elbow joint (26); and activating one or more motorised actuators (42, 44) to control movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) based on the measured torque in the respective shoulder joint (24) and / or elbow joint (26).

26. The method of claim 25, further comprising: estimating a voluntary joint torque in the shoulder joint (24) and / or elbow joint (26) based on the measured torque; and activating the one or more motorised actuators (42, 44) to control movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) based on the voluntary joint torque in the respective shoulder joint (24) and / or elbow joint (26).

27. The method of claim 26, further comprising determining an output torque or velocity for the one or more motorised actuators (42, 44) based on the voluntary joint torque, wherein activating the one or more motorised actuators (42, 44) to control movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) based on the voluntary joint torque in the respective shoulder joint (24) and / or elbow joint (26) comprises activating the one or more motorised actuators (42, 44) to control movement of the exoskeleton (14) about the shoulder joint (24) and / or elbow joint (26) according to the determined output torque or velocity.

28. The method of claim 27, wherein the output torque or velocity is proportional to the voluntary joint torque.

29. The method of claim 27 or 28, wherein, if the magnitude of the voluntary joint torque is below a first predetermined threshold, the output torque or velocity is zero.

30. The method of claim 29, wherein, if the magnitude of the voluntary joint torque is above a second predetermined threshold higher than the first predetermined threshold, the output torque or velocity is zero.

31. The method of any of claims 26 to 30 further comprising:using one or more sensors (40), measuring a bending moment or transferred shear force in the forearm portion (22) and / or upper arm portion (20); estimating an external weight acting on the device (10) based on the measured bending moment or shear force in the forearm portion (22) and / or upper arm portion (20); and estimating the voluntary joint torque additionally based on the estimated external weight.

32. The method of any of claims 26 to 31 , further comprising: defining a gravity model defining a torque contribution from at least one of a weight of the device and external weights acting on the device (10); and estimating the voluntary joint torque additionally based on the gravity model.

33. The method of claim 32, further comprising: using one or more sensors (49) to determine relative positions of the shoulder joint (24) and elbow joint (26) with respect to each other, and / or relative angular positions of the upper arm portion (20) and the forearm portion (22) with respect to the gravity vector; and using the relative positions and / or relative angular positions to define the gravity model.

34. The method of any of claims 26 to 33, further comprising: using one or more sensors (47), measuring an acceleration of the device (10); and estimating the voluntary joint torque additionally based on the measured acceleration.

35. The method of any of claims 25 to 34, further comprising adjusting control settings of the device (10) via an external communication device in wired or wireless communication with the device (10).

36. The method of claim 35, wherein the control settings comprise torque thresholds for controlling activation of the one or more motorised actuators (42, 44).

Images