Method for operating a human-machine interface

Abstract

Method for operating a human-machine interface (10) comprising the steps: - Applying a large number of accelerometers (12) to the skin (16) of a user by means of a holding device (14) - Calculating a user-requested input command from the user's muscle movements detected by the accelerometers (12), by the following procedure steps: ◯ Determining a common movement of the accelerometers (12) caused by a movement of the entire body part (18) of the user to which the accelerometers (12) are attached, ◯ Removing the accelerations of the accelerometers (12) caused by this movement of the accelerometers (12), so that only their relative accelerations to each other remain and are taken into account for further processing.

Description

[0001] The invention relates to a method for operating a human-machine interface.

[0002] Human-machine interfaces are known from the prior art and are used, for example, to control computers, prostheses, or robotic exoskeletons. In principle, a human-machine interface can be used to input any command to the control system of a device. It is known from the prior art to use electromyographic methods as human-machine interfaces, which allow the activation of individual muscles in a user's body part to be detected. Surface electromyography, in which the electrodes are applied to the skin, is a known method for this purpose.

[0003] It is also known to use pressure sensors that are pressed onto the skin, so that the pressure sensor registers when an underlying muscle tenses up.

[0004] Surface electromyography (SEM) faces the challenge of poor skin contact, which can change over time due to factors such as perspiration or user movement. Furthermore, signal acquisition and processing of very small potentials in the microvolt to millivolt range is extremely difficult under high-impedance contact (many kiloohms to several megaohms).

[0005] The use of pressure sensors has the disadvantage that these sensors must be pressed relatively firmly against the skin, which can become uncomfortable for the user over time. Furthermore, each individual sensor must be supported as stably and over as large an area as possible in order to deliver measurements.

[0006] This requires a very rigid mounting of the sensors, which limits convenience and ease of use.

[0007] Background information on the state of the art can be found in the following publications: US 2015 / 0 325 202 A1 JP 2009 - 95 431 A US 2015 / 0 084 860 A1

[0008] The object of the invention is to provide a method for operating a human-machine interface that delivers accurate measurements while also being user-friendly. It is preferred that at least one accelerometer is used per muscle to be detected. In most cases, several accelerometers are used per muscle to be detected.

[0009] The problem is solved by the features of claim 1.

[0010] An unclaimed human-machine interface on which the claimed method can be operated comprises a plurality of accelerometers and at least one holding device that holds the accelerometers to the user's skin. Indirect contact with the user's skin may also occur, for example via an intermediate element through which accelerations of the user's skin surface are transmitted to the accelerometers.

[0011] The human-machine interface also includes a processor for calculating a user-requested input command from the user's muscle movements detected by the accelerometers.

[0012] Using accelerometers to detect the user's muscle movement is advantageous because it can operate independently of contact difficulties, such as those encountered in surface electromyography. At the same time, it eliminates the need to press the sensors firmly against the user's skin. This results in more accurate sensor readings while simultaneously improving user-friendliness.

[0013] It is preferred that the holding device be an adhesive used to hold the user's accelerometers.

[0014] Alternatively, the holding device can be tubular or stump-shaped, so that it can be pulled over a part of the user's body, for example over the arm stump of a hand amputee, so that input commands for a prosthesis can be generated, for example.

[0015] Furthermore, it is preferred that the holding device is designed such that the acceleration sensors are movable relative to each other within a tolerance range. This is necessary because acceleration sensors must, of course, be movable to detect accelerations. This eliminates the need for a particularly rigid mounting of the sensors, thus increasing user-friendliness.

[0016] In a preferred embodiment, the acceleration sensors are one-dimensional sensors. However, they can also be multi-dimensional sensors, in particular three-dimensional sensors or inertial measurement units (IMUs). Investigations by the applicant have shown that accelerations of more than one thousandth of the acceleration due to gravity would be expected.

[0017] It is preferred that the human-machine interface for capturing the user's muscle movement uses only accelerometers as sensors and no other sensors.

[0018] The invention relates to a method for operating a human-machine interface, in particular as previously described. The method comprises the following steps: - Applying a large number of accelerometers to a user's skin using a holding device - Calculating a user-requested input command from the user's muscle movements detected by the accelerometers.

[0019] Furthermore, it is preferred that only accelerometers be used to detect the user's muscle movement. Alternatively, it is also possible to perform sensor fusion with other methods, such as electromyography or force measurement.

[0020] According to the invention, a common movement of the accelerometers is determined, which is caused by a movement of the entire body part of the user to which the accelerometers are attached. For example, it is possible that the user is intentionally moving their arm, and thus all accelerometers attached to this arm are moved along with it. The accelerations of the sensors caused by this movement are removed, so that only their relative movements to each other remain and are taken into account for further processing. The aforementioned process steps significantly improve the measurement accuracy.

[0021] A preferred embodiment of the invention is explained below with reference to a figure.

[0022] The figure schematically illustrates the structure of a human-machine interface.

[0023] The human-machine interface 10, as shown in the figure, is equipped with a tubular holding device 14 that is pulled over the user's arm stump 18. This holding device 14 holds the accelerometers 12 against the user's skin surface 16. It is preferred to use as many sensors as possible, for example, more than six one-dimensional accelerometers or more than two three-dimensional accelerometers. The movement of the muscles and tendons in the user's body part causes movement of the overlying skin 16, which is detected by the accelerometers 12. Desired input commands can be derived from this, for example, using detailed models of the forearm, including the muscles, known from the prior art. Furthermore, machine learning systems can be used that can map recurring patterns to specific actions.

[0024] The holding device 14 can, for example, be designed as a flexible and stretchable fabric.

[0025] Preferably, the sensors can detect contact with the environment, thus enabling additional input options. For example, it is possible to detect when a user taps their forearm, thereby generating a specific input command. A keypad can also be integrated into the holding device 14. Input commands can also be captured via knock codes or other forms of free touch.

[0026] Furthermore, it is possible to detect if several sensors are trapped between the user's body part 18 and a fixed object, such as a table. This would result in slightly altered patterns in the remaining sensors. Detecting such a condition can be helpful in ruling out misinterpretations.

Claims

[1] Method for operating a human-machine interface (10) comprising the steps: - Applying a large number of accelerometers (12) to the skin (16) of a user by means of a holding device (14) - Calculating a user-requested input command from the user's muscle movements detected by the accelerometers (12), by the following procedure steps: ◯ Determining a common movement of the accelerometers (12) caused by a movement of the entire body part (18) of the user to which the accelerometers (12) are attached, ◯ Removing the accelerations of the accelerometers (12) caused by this movement of the accelerometers (12), so that only their relative accelerations to each other remain and are taken into account for further processing. [2] Method according to claim 1, characterized by, that accelerometers (12) are used exclusively as sensors to detect the user's muscle movements. [3] Method according to one of claims 1 and 2, characterized by , that additional input commands are detected by external contact with the accelerometers (12) and / or the holding device (14), in particular by another part of the user's body.

Images