Reservoir system

Abstract

The reservoir system according to this embodiment has a plurality of sensors (1) and a reservoir (2). Each of the plurality of sensors (1) senses a time-series signal (S1) and is configured to send a time-series sensor signal (S2) based on the time-series signal (S1) to the reservoir (2). The reservoir (2) non-linearly transforms the sensor signal (S2) from each of the plurality of sensors (1). The reservoir (2) is configured to output output signals (S3, S4) as future prediction data before the input of the time-series sensor signal (S2) to the reservoir (2) is completed.

Description

Technical Field

[0001] The present disclosure relates to a reservoir system.

Background Art

[0002] A neuromorphic device is a device that mimics the human brain using a neural network. The neuromorphic device artificially mimics the relationship between neurons and synapses in the human brain.

[0003] A neuromorphic device has, for example, nodes (neurons in the brain) arranged in a hierarchical manner and transmission means (synapses in the brain) connecting between them. The neuromorphic device increases the correct answer rate of problems by the transmission means (synapses) learning. Learning is finding knowledge that may be useful in the future from information, and in a neuromorphic device, the input data is weighted.

[0004] As one type of neural network, a recurrent neural network is known. A recurrent neural network can handle non-linear time-series data. Non-linear time-series data is data whose value changes over time, and stock prices are an example of this. A recurrent neural network can perform data processing including past information by returning the processing result at the neurons in the subsequent layer to the neurons in the previous layer.

[0005] Reservoir computing is one means of realizing a recurrent neural network. Reservoir computing performs recursive processing by interacting signals based on internal connections. For example, Patent Document 1 discloses an information processing apparatus including a reservoir.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

SUMMARY OF THE INVENTION

PROBLEM TO BE SOLVED BY THE INVENTION

[0007] A neural network including reservoir computing predicts an estimated solution based on the input signal after the input of the input signal is completed. However, waiting for the input of the signal to be completed takes time to predict the estimated solution. This time lag inhibits timely prediction. Although the time lag can be shortened by accelerating the arithmetic processing using a high-performance computer, the system using a high-performance computer is at the cost of versatility.

[0008] The present disclosure has been made in view of the above circumstances, and provides a reservoir system capable of timely prediction.

MEANS FOR SOLVING THE PROBLEM

[0009] The reservoir system according to the first aspect has a plurality of sensors and a reservoir. Each of the plurality of sensors senses a time-series signal and is configured to send a time-series sensor signal based on the time-series signal to the reservoir. The reservoir non-linearly transforms the sensor signals from each of the plurality of sensors. The reservoir is configured to be able to output an output signal as future prediction data before the input of the time-series sensor signals to the reservoir is completed.

EFFECTS OF THE INVENTION

[0010] The reservoir system according to the above aspect enables timely prediction.

BRIEF DESCRIPTION OF THE DRAWINGS

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Embodiments for Carrying Out the Invention

[0012] Hereinafter, this embodiment will be described in detail with appropriate reference to the drawings. The drawings used in the following description may show, for the sake of clarity, the characteristic parts enlarged for the convenience of understanding the features of the present disclosure, and the specific configurations of each component may be different from the actual ones. The configurations and the like exemplified in the following description are merely examples, and the present disclosure is not limited thereto, and it can be implemented with appropriate modifications within the scope of achieving the effects of the present disclosure.

[0013] "First Embodiment" FIG. 1 is a block diagram of a reservoir system 100 according to this embodiment. The reservoir system 100 includes a plurality of sensors 1 and a reservoir 2.

[0014] Each of the plurality of sensors 1 is connected to the reservoir 2. Each of the sensors 1 senses a time-series signal S1 as an external signal. Each of the sensors 1 is configured to send a time-series sensor signal S2 based on the time-series signal S1 to the reservoir 2.

[0015] The type of the sensor 1 is not particularly limited. For example, an image detection sensor, an acceleration sensor, a temperature sensor, or an inertial sensor can be used as the sensor 1. The image detection sensor may be, for example, one that can represent a predetermined point (for example, a red part) in an image in coordinates. The acceleration sensor may be, for example, one that can be attached to a finger, an arm, etc. and can detect the movement of the arm with acceleration. Each of the sensors 1 may be of the same type or different types of sensors.

[0016] The time-series signal S1 is a signal that contains information on the measurement target that changes over time. For example, when sensor 1 is an acceleration sensor, the time-series signal S1 indicates the time change of the acceleration in a certain direction. For example, when sensor 1 is an image detection sensor, the time-series signal S1 indicates the time change of the coordinates of the measurement point.

[0017] Sensor 1 may output the time-series signal S1 as it is as the sensor signal S2. In this case, the time-series signal S1 coincides with the sensor signal S2. If the time-series signal S1 is used as the sensor signal S2 as it is, the information contained in the time-series signal S1 can be utilized without omission. Also, there is no time lag in converting the time-series signal S1 into the sensor signal S2, and the prediction of the reservoir system 100 can be performed in a timely manner.

[0018] Alternatively, sensor 1 may convert the time-series signal S1 and output it as the sensor signal S2. For example, sensor 1 may use, as the sensor signal S2, a time-division signal obtained by time-dividing the time-series signal S1 at regular intervals. By time-dividing the time-series signal S1, the computational load on the reservoir 2 can be reduced.

[0019] For example, sensor 1 may have a clock and a switch. The clock switches the ON / OFF of the switch based on a clock signal. The period of the clock signal may be, for example, 1 kHz or 1 MHz. Sensor 1 outputs the sensor signal S2 at regular intervals based on the clock signal. By operating sensor 1 based on the clock signal, the time-series signal S1 is converted into the sensor signal S2.

[0020] Alternatively, sensor 1 may further have a signal holding unit that holds a signal for a certain period. The signal holding unit holds the signal sensed by sensor 1 while the switch is switched by the clock signal. By holding the signals input during a certain period by the signal holding unit, the time-series signal S1 sensed during a certain period can be collectively output as the sensor signal S2. The signal holding unit is, for example, a capacitor.

[0021] For example, when each of the sensors 1 is a different sensor, the operating cycles of the respective sensors may be different. For example, depending on the type of sensor 1, the operating cycle may be faster than that of other types of sensors. By using a clock, it is possible to synchronize the periods of the sensor signals S2 input from each of the sensors 1. By synchronizing the periods of the sensor signals S2, it is possible to avoid a signal depending on one sensor 1 (for example, a signal with a fast operating cycle) from being input to the reservoir 2.

[0022] The reservoir 2 is connected to each of the plurality of sensors 1. The sensor signals S2 input from the plurality of sensors 1 into the reservoir 2 are non-linearly transformed within the reservoir 2.

[0023] The reservoir 2 may be software, hardware, or a combination of software and hardware.

[0024] Software is a program implemented in a computer. When the reservoir 2 is software, the reservoir 2 may be a microcontroller unit (MCU) or a microprocessor unit (MPU) including a central processing unit (CPU). The reservoir 2 includes, for example, a memory for storing a reservoir computing program and a processor for executing the reservoir computing program.

[0025] Hardware consists of a combination of actual elements or circuits. The hardware reservoir is referred to as a physical reservoir. The physical reservoir consists of a combination of actual elements or circuits and realizes the concept of reservoir computing with actual elements or circuits. For example, the physical reservoir may be a physical circuit. The physical reservoir may be, for example, an optical element using light, an electrical element using electricity, a magnetic element using magnetism, or a mechanical element using vibration. When the physical reservoir is an analog interface, it may have an analog-to-digital converter for digitally converting the output from the physical reservoir. Also, the physical reservoir may be implemented in a PLD (Programmable Logic device) such as an FPGA (Field-Programmable Gate Array).

[0026] Reservoir 2 includes, for example, a reservoir layer 21 and an output layer 22. The reservoir layer 21 is connected to a plurality of sensors 1. Also, the reservoir layer 21 is connected to the output layer 22.

[0027] The reservoir layer 21 stores the sensor signal S2 and converts it into another signal. The reservoir layer 21 sequentially converts the time-series sensor signal S2 without waiting for all of the time-series sensor signals S2 to be input. The reservoir layer 21 converts the data input to Reservoir 2 among the time-series sensor signals S2.

[0028] The reservoir layer 21 has a plurality of nodes n. The node n corresponds to a neuron in a neural circuit, and the connection between the nodes n corresponds to a synapse in a neural circuit. The node n non-linearly converts the input and outputs it. The number of nodes n does not matter. Each of the nodes n may be fully connected to a plurality of sensors 1. Each of the nodes n may be fully connected to the output layer 22.

[0029] Each node n is connected to one or more other nodes n. Each node n may be connected to all other nodes n within the reservoir layer 21, or may be connected to some of the nodes n within the reservoir layer 21. The connections between nodes n are random.

[0030] The connections between nodes n include recursive connections. A recursive connection is a connection where the output returns to the input. For example, a signal output from one node n at time t may return to the node n that output the signal after time t+1. This occurs when a signal output from one node n propagates through other nodes n and returns to the original node n. Thus, the connection relationship between nodes n where the output from a certain node n is input again via other nodes n is called a recursive connection. The reservoir 2 enables inference based on past information.

[0031] Connection coefficients indicating connection weights are set between nodes n. The signal input to a node n propagates between nodes n. The signal propagated to a certain node n is non-linearly transformed by the activation function of the node n, and further multiplied according to the connection coefficient, and then propagated to the next node n. The connection coefficients between nodes n can be arbitrarily set in the range from -1.0 to +1.0. The connection coefficients between nodes n are set by, for example, random numbers.

[0032] The signal exchange between nodes n may be an analog signal. An analog signal can handle time-series information and can perform time-series prediction of the reservoir system 100. Also, if there is no need to convert the analog signal to a digital signal, the power consumption of the entire reservoir system 100 can be reduced.

[0033] Also, the signal exchange between nodes n may be controlled using a clock. For example, each node n may have a clock and a switch. Each node n sends a signal to other nodes at a fixed timing based on the clock signal. By each node n operating based on the clock signal, the time-series signal S1 is converted into the sensor signal S2. Each node n may further have a signal holding unit that holds a signal for a certain period. The signal holding unit holds the signal received by node n while the switch is switched by the clock signal. By holding the signals input during a fixed period in the signal holding unit, the signals sensed during a fixed period can be grouped and sent to the next node n. By clock-controlling the signal exchange between nodes n, the timing of signal propagation between each node n can be synchronized.

[0034] The output layer 22 is connected to the reservoir layer 21. A coupling coefficient (weight) is set between each node n of the output layer 22 and the reservoir layer 21. This coupling coefficient is updated and optimized during the learning of the reservoir system 100. During inference using the reservoir system 100, this coupling coefficient is fixed according to the learning result.

[0035] The output layer 22 applies a weight to the output signal S3 from the reservoir layer 21. Applying the weight corresponds to a process of performing a product operation on the output signal S3 with the coupling coefficient. The process of the output layer 22 may be performed as a software arithmetic process or, for example, a process using hardware such as a multiply-accumulate circuit.

[0036] The output layer 22 outputs an output signal S4 based on the output signal S3 from the reservoir layer 21. The output signal S3 is a conversion of the sensor signal S2 input to the reservoir 2 and has part of the information of the signal input to the reservoir 2. The output signal S4 is output based on the signal input to the reservoir 2.

[0037] The output layer 22 outputs the output signal S4 before the input of the time-series sensor signal S2 to the reservoir 2 is completed. That is, the timing at which the output signal S4 is output is earlier than the timing at which the input of the sensor signal S2 is completed. The output layer 22 outputs the output signal S4 during the input of the sensor signal S2. The output layer 22 outputs the output signal S4 using the past data input to the reservoir 2 by the time the output signal S4 is output among the time-series sensor signals S2. That is, the reservoir 2 can predict future data that has not yet been input from the past data and output the output signal S4.

[0038] For example, when the output layer 22 is software, the output layer 22 is programmed to output the inference result based on the current information before the input of the sensor signal S2 to the reservoir 2 is completed, without waiting for the input.

[0039] For example, when the output layer 22 is hardware, the output layer 22 has a signal propagation path that does not pass through a capacitor or the like that stores all the information of the sensor signal S2 until the input to the reservoir 2 is completed. By using this signal propagation path, the output layer 22 outputs the inference result based on the current information as the output signal S4 before the input of the sensor signal S2 to the reservoir 2 is completed.

[0040] The output signal S4 may change with the passage of time. The output signal S4 may be, for example, an analog signal. The output signal S4 may be a digital signal obtained by digitally converting an analog signal. The output signal S4 may be output mainly over time in multiple times. For example, when the sensor signal S2 is a time-division signal, the number of times the output signal S4 is output may be less than or equal to the number of sampling times at which the time-division sensor signal S2 is input to the reservoir 2.

[0041] The output signal S4 is output from the output unit, for example, in a form recognizable by the user. The output unit may be, for example, an image display device such as a liquid crystal display or an organic EL (Electro Luminescence) display. The output unit may also be an interface for connecting to an image display device. In this case, the output unit generates a video signal for displaying image data and outputs the video signal to the image display device connected to itself. The output unit may be a device that outputs sound such as a speaker. The output unit may also be an interface for connecting to a sound output device such as a speaker or headphones.

[0042] Next, the operation of the reservoir system 100 according to the first embodiment will be described. The reservoir system 100 can perform learning and inference.

[0043] First, the learning of the reservoir system 100 will be described. When performing the learning of the reservoir system 100, the output signal S4 and the teacher signal are compared, and the weight (coupling coefficient) applied to the output signal S3 is adjusted so that the degree of coincidence (mutual information amount) between the output signal S4 and the teacher signal becomes high.

[0044] The teacher signal may be stored in the storage unit, for example. The storage unit is configured using a storage medium such as a magnetic hard disk device or a semiconductor storage device. The storage unit may also be a computer-readable recording medium. A computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, a semiconductor storage device (e.g., SSD: Solid State Drive), or a storage device such as a hard disk or a semiconductor storage device built into a computer system.

[0045] A specific example will be given for explanation. For example, when the sensor 1 is an image detection sensor and the reservoir system 100 is a system that predicts the number drawn by a finger from the movement of the finger. For example, a red finger sack is attached to the finger, and the image detection sensor detects the red point in the image.

[0046] FIG. 2 is a diagram for explaining the operation of the reservoir system 100 according to the first embodiment. As shown in FIG. 2, when a finger draws the numerical value "2", the movement of the finger can be converted into the transition of the coordinates of the X-axis and the Y-axis. Each of the movement of the X-axis and the movement of the Y-axis becomes the time-series signal S1 detected by the sensor 1.

[0047] In this case, the reservoir system 100 should output the output signal S4 of "2". That is, the solution of "2" becomes the teacher signal. When the reservoir system 100 does not output the solution of "2", the weight (coupling coefficient) applied to the output signal S3 is adjusted. Such processing is performed for various numbers and the finger movements of various subjects to optimize the weight (coupling coefficient) applied to the output signal S3.

[0048] Another specific example will be described. For example, an example is illustrated in the case where the sensor 1 is an acceleration sensor and the reservoir system 100 is a system that predicts the shape of the fingers of a hand. For example, an acceleration sensor is installed on each of the five fingers, and the acceleration sensor detects the movement of the fingers.

[0049] For example, the movement of the finger is different between the case of grasping the finger and the case of spreading the finger, and the signals detected by the acceleration sensor are also different. For example, when the operation of grasping the finger is performed, it is correct for the reservoir system 100 to output the output signal S4 of "grasp the finger" based on the time-series signal from the acceleration sensor. That is, the gesture of "grasp the finger" becomes the teacher signal. When the reservoir system 100 does not output the gesture of "grasp the finger", the weight (coupling coefficient) applied to the output signal S3 is adjusted. Such processing is performed for various numbers and the finger movements of various subjects to optimize the weight (coupling coefficient) applied to the output signal S3.

[0050] Next, the inference of the reservoir system 100 will be described. When the reservoir system 100 performs inference, the sensor 1 senses an external signal as a time-series signal S1. The time-series signal S1 becomes a sensor signal S2 and is input to the reservoir 2. Inside the reservoir 2, the sensor signal S2 becomes an output signal S3 and weights are applied. The weights applied at this time are the weights determined during learning. Then, the reservoir system 100 outputs an output signal S4 based on the output signal S3. The output signal S4 is output as future prediction data before the input of the time-series sensor signal S2 to the reservoir 2 is completed.

[0051] A specific example will be given for explanation. For example, a case where the sensor 1 is an image detection sensor and the reservoir system 100 is a system that predicts the number drawn by a finger from the movement of the finger will be illustrated.

[0052] FIG. 3 is a diagram for explaining the operation of the reservoir system 100 according to the first embodiment. The movement when a finger draws a numerical value is input to the reservoir system 100 as a time-series signal S1. The time-series signal S1 is a signal including all information from the start to the end of writing the numerical value. The reservoir system 100 starts prediction simultaneously with the start of the input of the time-series signal S1.

[0053] For example, at the 10-frame time point, the number has just started to be written and the input of the time-series signal S1 is not completed. For example, at the 20-frame time point, the number is being written and the input of the time-series signal S1 is not completed. For example, it is not until the 42-frame time point that the input of the time-series signal S1 is completed.

[0054] The reservoir system 100 outputs an estimated solution at each stage. For example, at the 5-frame time point, an estimated solution that there is a possibility of drawing '7' or '8' is output. For example, at the 20-frame time point, an estimated solution that there is a possibility of drawing '7' is output. For example, at the 42-frame time point, an estimated solution that it was highly likely to draw '2' is output.

[0055] For example, when drawing the number "2", there is a characteristic bend just before drawing the last horizontal line. Such a movement is characteristic of "2" or "3", and at around the 25 - frame mark, the reservoir system 100 outputs an estimated solution with the highest probability of "2".

[0056] The reservoir system 100 according to the first embodiment was predicting the movement of the finger about 20 frames earlier than the numerical value was finished being drawn at 42 frames. The reservoir system 100 according to the first embodiment can make timely predictions and can reduce the time lag.

[0057] Here, the reservoir system 100 outputs inappropriate estimated solutions before 20 frames. This is also something that is done in normal human judgment and is not a problem. For example, even when a human tries to predict the numerical value being drawn while the number "2" is being drawn, the same thing happens. Even if it is not correct, making a timely prediction can increase the speed of later judgment.

[0058] Here, a specific example of a system for predicting hand movement using an image detection sensor has been described, but the same timely prediction can be made in the case of other sensors as well.

[0059] For example, even in the case of a system that predicts the shape of the fingers of the hand by attaching an acceleration sensor to the finger or arm, from the information on the change in the acceleration of the finger or arm before the actual movement such as "grasping the finger" or "spreading the finger" is completed, the shape of the fingers of the hand can be predicted. For example, if the shape of the fingers of the hand can be predicted, in rock - paper - scissors, before the opponent actually shows the shape of the hand, it is possible to predict what the opponent will show.

[0060] Here, an example where learning and inference are performed at different timings has been shown, but learning and inference may also be performed in parallel. For example, learning may be performed by online learning. Online learning updates the coupling coefficient every time one piece of data is input. By performing online learning, even when the response changes with time, the accuracy of the estimated solution can be improved.

[0061] "Second Embodiment" FIG. 4 is a schematic diagram of a reservoir system 101 according to the second embodiment. The reservoir system 101 includes a plurality of sensors 1, a reservoir 2, an input detection unit 3, and an output detection unit 4. The reservoir system 101 is different from the reservoir system 100 in that it has an input detection unit 3 and an output detection unit 4. In the reservoir system 101, the same components as those in the reservoir system 100 are denoted by the same reference numerals, and the description thereof is omitted.

[0062] The input detection unit 3 is connected to each of the plurality of sensors 1 and the reservoir 2. The input detection unit 3 monitors the sensor signal S2 and is configured to be able to determine the start point of the signal used for predicting the prediction data among the sensor signals S2.

[0063] The input detection unit 3 may be software, hardware, or a combination of software and hardware.

[0064] The input detection unit 3 may be, for example, a microcontroller unit (MCU) or a microprocessor unit (MPU) including a central processing unit (CPU), or an input detection circuit.

[0065] The input detection unit 3 determines, for example, the change point as the start point when the sensor signal S2 from the sensor 1 changes from a certain state. The degree of change from a certain state can be freely designed according to the task, application, etc. given to the reservoir system 101. For example, the input detection unit 3 may determine the time point when the change amount of the sensor signal S2 exceeds a certain threshold as the start point, or may determine the time point when the sensor signal S2 exceeds a certain threshold as the start point.

[0066] The output detection unit 4 is connected to the reservoir 2. The output detection unit 4 monitors the output signal S4 and is configured to be able to determine the end point of the signal used for predicting the prediction data among the sensor signals S2.

[0067] The output detection unit 4 may be software, hardware, or a combination of software and hardware.

[0068] The output detection unit 4 may be, for example, a microcontroller unit (MCU) or a microprocessor unit (MPU) including a central processing unit (CPU), or an output detection circuit.

[0069] For example, when the output signal S4 changes from a certain state, the output detection unit 4 determines the change point as the start point. The degree of change from a certain state can be freely designed according to the tasks, applications, etc. given to the reservoir system 101. For example, the output detection unit 4 may determine the point in time when the change amount of the output signal S4 exceeds a certain threshold as the start point, or may determine the point in time when the output signal S4 exceeds a certain threshold as the start point.

[0070] Also, the output detection unit 4 may be between each of the plurality of sensors 1 and the reservoir 2. When the sensor signal S2 changes from a certain state, the output detection unit 4 may determine the change point as the end point. For example, the output detection unit 4 may determine the point in time when the change amount of the sensor signal S2 becomes equal to or less than a certain amount as the end point.

[0071] Similar to the reservoir system 100 according to the first embodiment, the reservoir system 101 according to the second embodiment can perform timely prediction. Also, since the reservoir system 101 can set the start point and end point of the prediction, there is no need to perform unnecessary calculations, and the power consumption of the reservoir system 101 can be reduced.

[0072] In the reservoir system 101, an example having both the input detection unit 3 and the output detection unit 4 is shown, but only one of them may be used. Even just setting the start point or the end point can sufficiently reduce the calculation load.

[0073] As described above in detail with reference to the drawings for several embodiments, each configuration and their combinations in each embodiment are examples, and additions, omissions, substitutions, and other changes of the configuration are possible without departing from the spirit of the present disclosure.

Description of Reference Numerals

[0074] 1 Sensor 2 Reservoir 3 Input Detection Unit 4 Output Detection Unit 21 Reservoir Layer 22 Output Layer 100, 101 Reservoir System n Node S1 Time-Series Signal S2 Sensor Signal S3, S4 Output Signal

Claims

1. a plurality of sensors and a reservoir; each of the plurality of sensors is configured to sense a time series signal and transmit a time series sensor signal based on the time series signal to the reservoir; the reservoir nonlinearly transforms a sensor signal from each of the plurality of sensors; A reservoir system, wherein the reservoir is configured to output an output signal as future predicted data before input of the time series of the sensor signal into the reservoir is complete.

2. The reservoir system according to claim 1 , wherein the output signal is determined using past data of the time series of the sensor signal that was input to the reservoir before the output signal was output.

3. Each of the plurality of sensors inputs a time series signal into the reservoir in a time-division manner at a fixed time interval; the reservoir outputs the output signal multiple times over time; 2. The reservoir system according to claim 1, wherein the number of times the output signal is output from the reservoir is equal to or less than the number of times the time-division time series signal is sampled and input to the reservoir.

4. The reservoir system of claim 1 , wherein the sensor signal corresponds to the time series signal.

5. The reservoir system according to claim 1 , wherein the sensor signal is a signal obtained by time-dividing the time series signal at regular intervals.

6. the reservoir comprises a reservoir layer and an output layer; the reservoir layer converts the time series of sensor signals input to the reservoir; 2. The reservoir system of claim 1, wherein the output layer applies weights to the outputs from the reservoir layer and outputs the output signals based on the signals input to the reservoirs.

7. The reservoir system of claim 1 , wherein the output signal is an analog signal.

8. the reservoir having a plurality of nodes; The reservoir system of claim 1 , wherein signal exchange between the plurality of nodes is performed using analog signals.

9. the reservoir having a plurality of nodes; The reservoir system of claim 1 , wherein each of the plurality of nodes transmits a signal to the other nodes at a determined period.

10. The reservoir system of claim 1 , wherein at least one of the plurality of sensors is an image detection sensor.

11. The reservoir system of claim 1 , wherein at least one of the plurality of sensors is an acceleration sensor.

12. An input detection unit is further provided, the input detection unit is connected to each of the plurality of sensors and to the reservoir; The reservoir system of claim 1 , wherein the input detector is configured to monitor the sensor signal and determine a starting point of the sensor signal to be used to predict the predicted data.

13. Further comprising an output detection unit, the output detection unit is connected to the reservoir; 2. The reservoir system of claim 1, wherein the output detector is configured to monitor the output signal and determine an end point of the sensor signal that is utilized to predict the predicted data.

14. the plurality of sensors are acceleration sensors that can be worn on a finger or an arm, The reservoir system of claim 1 , wherein the reservoir is configured to predict hand movement as the output signal.

15. the plurality of sensors are sensors capable of indicating a predetermined point in an image by coordinates; The reservoir system of claim 1 , wherein the reservoir is configured to use the coordinate progression to predict hand movement as the output signal.

Images