JPH07175771A - Neural network and signal processor, autonomous system, autonomous robot and moving system using the same - Google Patents

Abstract

PURPOSE:To easily construct a complicated system by a simple system by connecting plural neuron elements of multiple inputs and one output where the displacement amount of an internal active value according to a neuron output value is determined with each other. CONSTITUTION:In an input/output relation of a neuron element 11, a multiple- input and one-output constitution that parallel input signals x1(t) to XN(t) from other neuron element or the outside are inputted and an output signal y(t) is outputted is taken. This neuron element 11 is provided with an adder 12 adding plural signals and a nonlinear converter 13 performing the nonlinear conversion for the addition result and outputting the result successively. Further, the input side of the adder 12 is connected with plural multipliers 14, an arithmetic unit 15 and a memory 16 storing a threshold theta. In the neuron element 11, the displacement amount of an internal active value u(t) is determined according to its own output value, and a neural network is constituted by connecting plural neuron element 11 to each other.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a neural network imitating a nerve cell, which can be applied to various controls such as recognition of images and voices, adaptive control of robots, associative memory, nonlinear prediction, etc. The present invention relates to a signal processing device, an autonomous system, an autonomous robot, and a mobile system used.

[0002]

2. Description of the Related Art The function of a nerve cell (neuron), which is a basic unit of information processing in a living body, is mimicked, and further, this "nerve cell mimicking element" (nerve cell unit) is networked to perform parallel processing of information. What we aimed for was a so-called neural network. Although it is easy to perform character recognition, associative memory, motion control, etc. in a living body, there are many things that conventional Neumann computers cannot easily achieve. Attempts have been actively made to solve these problems by imitating the neural system of a living body, in particular, the functions peculiar to the living body, that is, parallel processing, self-learning, etc. by a neural network.

However, most of the studies on these neural networks deal with static states, which are different from the dynamic states that are actually performed in the living body. Therefore, in order to realize the information processing function of humans and other living bodies, it is necessary to handle this dynamic state.

The "chaos phenomenon" is considered to play a major role in the information processing of dynamic neural networks. An example of the neural network model having this chaotic state is shown in "Chaos in Neural Systems" (edited by Kazuyuki Aihara, Tokyo Denki University Press, pp. 157-188).

FIG. 24 shows a block diagram configuration of an electronic circuit model of a chaotic neuron shown in this document. This uses two sample-and-hold circuits (S & H circuits) 1 and 2 and a one-dimensional mapping circuit 3, and an S & H circuit 1 holds an internal state value one step before according to a clock 4, and this is stored in a one-dimensional mapping circuit. 3 and the value of the next time is obtained at the output of the S & H circuit 2 in accordance with the clock 5 and the operation is repeated.

[0006]

However, the chaotic neuron model as shown in the above document is composed of an analog circuit as shown in FIG. Furthermore, although it is a simple model, it still has complexity, such as the existence of a refractory period. Therefore, the function of the dynamic neural network is not yet fully utilized.

[0007]

A neural network according to a first aspect of the present invention comprises an adding means for adding a plurality of signals,
Non-linear conversion means for performing non-linear conversion of the internal activation value resulting from the addition of the addition means and outputting the conversion result to the outside as a neuron output value, and an external input signal multiplied by a coupling coefficient and outputting the multiplication result to the addition means. And external input means for performing arithmetic processing on the internal activation value and the neuron output value to output to the adding means, and the displacement amount of the internal activation value is changed according to the neuron output value. A plurality of determined neuron elements with one input and one output are connected to each other.

According to a second aspect of the present invention, in addition to the structure of the second aspect of the neural network, a storage means is provided which is connected to the addition means and stores the internal activation value.

In the neural network according to a third aspect of the present invention, in the configuration of the neural network according to the second aspect, the storage means for storing the threshold value for determining the displacement amount of the internal activation value is also used as the storage means for the coupling coefficient. It is a thing.

According to a fourth aspect of the present invention, in addition to the structure of the second or third aspect of the neural network, latch means is provided on the output line from the external input means to the adding means.

According to a fifth aspect of the present invention, in the neural network structure according to the first, second, third or fourth aspect, the displacement amount of the internal activation value is calculated as follows: (the difference between the input value of the external input signal and the coupling coefficient value) Sum of products) + (threshold)
It is calculated as − (product of internal activation value and neuron output value).

According to a sixth aspect of the present invention, there is provided a signal processing device including the neural network according to the first, second, third, fourth and fifth aspects, and the input / output signals of the neural network as input to obtain a coupling coefficient, a threshold value, etc. And a parameter updating means by a genetic algorithm computing means for calculating a parameter and outputting it to the neural network.

A signal processing apparatus according to a seventh aspect of the present invention is a neural network according to the first, second, third, fourth or fifth aspect, and a parameter such as a coupling coefficient or a threshold value of the neural network with an input signal to the neural network as an input. And a parameter updating means by a genetic algorithm calculating means for calculating and outputting to the neural network.

A signal processing device according to claim 8 is the neural network according to claim 1, 2, 3, 4 or 5, and the input / output signals of the neural network are used as inputs to determine the coupling coefficient, threshold value, etc. of the neural network. It is constituted by a parameter updating means by a random search computing means for calculating a parameter and outputting it to this neural network.

According to a ninth aspect of the present invention, there is provided a signal processing apparatus including the neural network according to the first, second, third, fourth or fifth aspect, and a parameter such as a coupling coefficient or a threshold value of the neural network with an input signal to the neural network as an input. And a parameter updating means by a random search computing means for calculating and outputting to the neural network.

According to a tenth aspect of the present invention, there is provided a signal processing device according to the sixth, seventh, eighth or ninth aspect, wherein the signal processing device has a plurality of neural networks.

According to another aspect of the invention, an autonomous system comprises a neural network whose state dynamically changes, and signal input means for inputting a signal to the neural network.
And a signal output means for inputting a signal output from the neural network.

According to a twelfth aspect of the present invention, with respect to the configuration of the autonomous system according to the eleventh aspect, the neural network is the neural network according to the first, second, third, fourth or fifth aspect.

An autonomous system according to a thirteenth aspect of the present invention relates to the configuration of the autonomous system according to the eleventh aspect, wherein the neural network includes an addition means for adding a plurality of signals and a non-linear conversion means for nonlinearly converting the addition result of the addition means. An output means for outputting the conversion result by the non-linear conversion means to the outside; an external input means for multiplying a plurality of external input signals by a coupling coefficient and outputting the multiplication result to the adding means; and addition by the adding means. First internal input means for multiplying the result by a coefficient and outputting the multiplication result to the adding means, and second internal means for multiplying the conversion result of the non-linear converting means by the coefficient and outputting the multiplication result to the adding means. It is a chaotic neural network in which a plurality of neuron elements having an input means are connected to each other.

An autonomous system according to a fourteenth aspect is the autonomous system according to the thirteenth aspect, wherein each neuron element in the chaotic neural network has a storage means for storing a signal to be input to the addition means. Is.

According to a fifteenth aspect of the invention, in the configuration of the autonomous system according to the thirteenth or fourteenth aspects, the addition result of the adding means is directly output to the adding means as the first internal input means.

An autonomous system according to a sixteenth aspect is the autonomous system according to the thirteenth, fourteenth or fifteenth aspects, wherein the result of conversion by the non-linear conversion means is directly output to the addition means as the second internal input means. is there.

According to a seventeenth aspect of the present invention, in the configuration of the autonomous system according to the thirteenth, fourteenth, fifteenth or sixteenth aspect, the chaotic neural network generates an input signal and inputs an output result to input these signals. Evaluation means for generating an evaluation value from the signal and the output result,
Genetic coefficient calculation means is provided for reading out the coefficient inside the neuron element using the evaluation value by the evaluation means, updating the coefficient according to the genetic algorithm, and writing the updated coefficient to each neuron element.

An autonomous system according to claim 18 is the structure of the autonomous system according to claim 13, 14, 15 or 16 in which the input signal is used as an external input means for updating the coupling coefficient.

An autonomous robot according to a nineteenth aspect receives a neural network whose state dynamically changes, a sensor which generates a signal to be input to the neural network, and a signal output from the neural network as control signals. Drive display means for driving or displaying.

According to a twenty-first aspect of the invention, the neural network in the configuration of the autonomous robot according to the nineteenth aspect is the neural network according to the first, second, third, fourth or fifth aspect.

According to a twenty-first aspect of the invention, there is provided an autonomous robot, wherein the neural network in the configuration of the autonomous robot according to the nineteenth aspect is added by an adding means for adding a plurality of signals and a non-linear conversion for nonlinearly converting the addition result of the adding means. Means, an output means for outputting the conversion result by the non-linear conversion means to the outside, an external input means for multiplying a plurality of external input signals by a coupling coefficient and outputting a multiplication result to the adding means, and an adding means of the adding means. A first internal input means for multiplying the addition result by a coefficient and outputting the multiplication result to the adding means; and a second internal input means for multiplying the conversion result of the non-linear conversion means by the coefficient and outputting the multiplication result to the adding means. This is a chaotic neural network in which a plurality of neuron elements having internal input means are connected to each other.

In the autonomous robot according to claim 22, each neuron element in the chaotic neural network in the configuration of the autonomous robot according to claim 21 has a storage means for storing a signal to be inputted to the adding means. It is a thing.

An autonomous robot according to a twenty-third aspect of the present invention is the autonomous robot according to the twenty-first or twenty-second aspect, wherein:
This is the first internal input means for directly outputting the addition result of the adding means to this adding means.

The autonomous robot according to a twenty-fourth aspect is the second internal input means for directly outputting the conversion result of the non-linear conversion means to the addition means in the configuration of the autonomous robot according to the twenty-first, twenty-second or twenty-third aspect. It is a thing.

According to a twenty-fifth aspect of the present invention, the autonomous robot according to the twenty-first, twenty-second, twenty-third, or twenty-fourth aspect of the present invention generates an input signal and inputs an output result into the chaotic neural network in the configuration of the autonomous robot. An evaluation means for generating an evaluation value from the input signal and the output result, and a coefficient inside the neuron element is read using the evaluation value by this evaluation means, the coefficient is updated according to a genetic algorithm, and the updated coefficient is applied to each neuron element. And a genetic algorithm calculating means for writing.

According to a twenty-sixth aspect of the invention, the autonomous robot according to the twenty-first, twenty-second, twenty-third, or twenty-fourth aspect is an external input means for updating the coupling coefficient using an input signal.

According to a 27th aspect of the present invention, there is provided a moving system, which comprises a potential detecting means for detecting a potential of a work area.
The neural network according to claim 1, 2, 3, 4 or 5, which receives a detection output from the potential detecting means, a signal converting means for converting and outputting an output signal from the neural network, and the signal converting means. The steering wheel and the driving wheel are provided with output signals from the control signals.

According to a twenty-eighth aspect of the present invention, there is provided a moving system, which comprises potential detecting means for detecting a potential of a work area.
The signal processing device according to claim 6, 7, 8, 9 or 10, wherein the detection output by the potential detection means is input.
It is provided with a signal converting means for converting and outputting an output signal from the signal processing device, and steering wheels and driving wheels to which the output signal from the signal converting means is given as a control signal.

In the mobile system according to claim 29, the parameter updating means in the configuration of the mobile system according to claim 27 or 28 is provided outside the mobile body and is connected by communication means.

[0036]

In the neural network according to the first aspect of the present invention, since a plurality of multi-input / single-output neuron elements in which the displacement amount of the internal activation value is determined according to the neuron output value are connected to each other, a steady state or , It is possible to show dynamic states such as chaos, limit cycle, etc., and a complex system can be easily constructed with a simple network.

In the neural network according to the second aspect of the invention, since the internal activation value storage means is provided, a plurality of signals can be input to the same signal line in time series,
Expandability of the neural network according to claim 1,
The processing speed can be increased and the system flexibility is increased.

In the neural network according to the third aspect, the storage means for storing the threshold value for determining the displacement amount of the internal activation value is also used as the storage means for the coupling coefficient. Therefore, the neural network according to the second aspect. The scale of neuron elements in the network can be reduced, and the scale of the neural network can be increased.

In the neural network according to the fourth aspect, since the latch means is provided on the output line from the external input means to the adder means, pipeline processing becomes possible, processing speed is improved, and system flexibility is increased. Becomes

In the neural network according to the present invention, the amount of displacement of the internal activation value is calculated by (sum of products of input value of external input signal and coupling coefficient value) + (threshold value)-(internal activation value and neuron output). Since it is calculated as a product of values, the neural network according to claim 1, 2, 3 or 4 can be easily realized.

In the signal processing device according to claim 6 or 7, the input / output signal or the input signal of the neural network is input to the neural network according to claim 1, 2, 3, 4 or 5 as described above. Since the parameter updating means by the genetic algorithm calculating means for calculating the parameters such as the coupling coefficient and the threshold value of the neural network and outputting them to the neural network is combined, the parameter updating of the neural network becomes possible and the applicability of the system is enhanced.

Also in the signal processing device according to claim 8 or 9, the input / output signal or the input signal of this neural network is input to the neural network according to claim 1, 2, 3, 4 or 5 as described above. Since the parameter updating means by the random search calculation means for calculating the parameters such as the coupling coefficient and the threshold value of the neural network and outputting them to this neural network is combined, the parameter updating of the neural network becomes possible and the applicability of the system is enhanced. As compared with the case of using the genetic algorithm calculating means, the parameter updating means is simpler and hardware implementation is easier.

Since the signal processing apparatus according to the tenth aspect has a plurality of neural networks, the processing speed of the signal processing apparatus according to the above-described sixth, seventh, eighth or ninth aspect is improved, and a plurality of neural networks are provided. It is also possible to improve performance by allowing the input and output signals of the network to interact with each other.

In the autonomous system according to the eleventh to eighteenth aspects, since the autonomous system is configured by using the neural network whose state dynamically changes, it is possible to construct a highly autonomous system.

In the autonomous robot according to any one of claims 19 to 26, since the autonomous robot is constituted by using the neural network whose state dynamically changes, it is possible to provide a highly autonomous robot.

In the mobile system according to the twenty-seventh to twenty-ninth aspects, the mobile system is configured by using a neural network whose state dynamically changes or a signal processing device equipped with such a neural network.
With a very simple structure, it is possible to build a complicated mobile system.

[0047]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the invention described in claims 1 and 5 is shown in FIG.
Or, it demonstrates based on FIG. FIG. 1 and FIG. 2 show a configuration example of one neuron element 11 which is an element constituting the neural network of this embodiment showing dynamic states such as periodic output and chaotic output due to internal or external factors. Show.

First, FIG. 2 shows the input / output relationship of one neuron element 11, and parallel input signals x ₁ (t) to x _N (t) (N is Any positive integer, t means time) and the output signal y
It has a multi-input / single-output configuration for outputting (t).

FIG. 1 is a block diagram showing a configuration for internal processing of such a neuron element 11. First, an adder (adding means) 12 for adding a plurality of signals and a non-linear converter (non-linear converting means) 13 for non-linearly converting and outputting the addition result are sequentially provided. To the input side of the adder 12, a plurality of multipliers 14 serving as external input means, a computing unit 15 serving as internal input means, and a memory 16 storing a threshold value θ are connected. The multiplier 14 calculates the product of each input signal x _i (t) (i = 0 to N) and the coupling coefficient w _i stored in the memory 17. Further, the arithmetic unit 15 has an internal activation value u (t) which is an output value of the adder 12.
And the output signal y of the non-linear converter 3 that is the neuron output value
Predetermined arithmetic processing is performed on (t) and. Therefore, in such a neuron element 11, the amount of displacement of the internal activation value u (t) is determined according to the output value of the own neuron element 11.

A concrete calculation method in such a neuron element 11 is represented by, for example, equations (1) to (3). However, in the case of a discrete system, equation (1) becomes equation (4).

[0051]

[Equation 1]

In these cases, the processing in the computing unit 15 is {1-y (t)} u (t), and the processing in the non-linear computing unit 3 is the equation (2). This corresponds to the arithmetic processing of the invention described in claim 5.

A plurality of such neuron elements 11 are prepared and connected to each other to form a neural network. For example, as shown in FIG. 3, three neuron elements 11A to 11C are prepared, the inputs and outputs are mutually connected, and, for example, an input signal from the outside is given to the neuron element 11A, and the neuron elements 11B and 11C are supplied. The outputs 1 and 2 are respectively taken out. With such an extremely simple system configuration, a chaotic behavior is shown when the input value is a certain value, and a periodic behavior is shown when the input value is a value, thus constructing a powerful nonlinear system. It is suggested that it can be done. Therefore, such a neural network is a neural network capable of showing a steady state and dynamic states such as chaos and limit cycle. That is, the output value of each neuron element 11 exhibits dynamic behavior such that the output values of the neuron elements 11 dynamically transition to each other due to an external input signal, and a neural network having adaptability to a dynamic state can be constructed. Become.

The non-linear conversion unit 3 may convert the output signal from the adder 12 using table lookup, or may calculate the non-linear function by a DSP or a CPU. Also, adder 1
2, the multiplier 14 and the arithmetic unit 15 may also be realized by a DSP or a CPU, and the memory 16 may be a register that holds only one data. Further, such a neuron element 11 may be partially or entirely realized by software.

An embodiment of the invention described in claim 2 will be described with reference to FIG. The same parts as those shown in the above-mentioned embodiments are designated by the same reference numerals (the same applies to the following embodiments). In the above-described embodiment, the neuron element 11 to which the input signal x _i (t) is input in parallel is configured, but in the present embodiment, the input signal x _i (t) is sequentially, that is, in time series. It is configured as an input neuron element 18. In comparison with FIG. 1, the adder 1 is connected between the input and output of the adder 12 and
A memory 19 for storing the internal activation value u (t) according to the output signal of 2 is added as a storage means. Also, the computing unit 1
An adder 10 for adding the calculation result of 5 and the threshold value θ stored in the memory 16 and outputting to the memory 19 is also added. Further, the input signal x _i (t) to the adder 12
Has a single signal line, the memory 17 holds all coupling coefficients w _i corresponding to the respective inputs, and the input signal x is controlled by an address controller (not shown).
It is configured to output the coupling coefficient w _i corresponding to _i (t) to the multiplier 14. Alternatively, the memory 17 may be composed of a shift register.

In such a configuration, the input signal x _i (t)
Are sequentially input, and the product with the corresponding coupling coefficient w _i held in the memory 14 is calculated in the multiplier 14. The output value of the multiplier 14 is added by the adder 12 with the sum of the input signals x ₁ (t) to x _i-1 (t) held in the memory 19, and nonlinearly converted by the nonlinear converter 3. After that, the output signal y (t) is output. The final sum u (t) of the adder 12 at the previous time (t-1)
-1) (sum after all input signals x _i (t-1) are input),
The output signal y (t-1) at the previous time (t-1) is arithmetically processed by the arithmetic unit 15, and is added by the threshold value θ held in the memory 16 and the adder 20, and the obtained sum is stored in the memory 19 Loaded as the initial value.

As in the case of the neuron element 11, a plurality of such neuron elements 18 are prepared and connected to each other in accordance with FIG. 3 to form a neural network. According to the neural network of the present embodiment, it is possible to input a plurality of input signals x _i (t) on the same line in time series.
Since the expandability as a neural network and the processing speed can be improved, the flexibility of the system is increased.

An embodiment of the invention described in claim 3 will be described with reference to FIG. In the present embodiment, in the configuration of the embodiment shown in FIG. 4, the memory 17 which stores and holds the coupling coefficient w _i also serves as a storage means which also stores the threshold value θ for determining the internal displacement amount u (t). Is configured. That is,
The threshold θ is also stored in the memory 17, and the threshold θ is stored in the memory 17.
If the multiplier 14 is such that the input signal becomes 1 when the threshold value θ is output, the threshold value θ is directly input from the multiplier 14 to the adder 12. Become. Therefore, the neuron element 21 of the present embodiment
Is the memory 16 from the neuron elements 18 shown in FIG.
And the adder 20 is omitted.

As in the case of the neuron element 11, a plurality of such neuron elements 21 are prepared and connected to each other according to FIG. 3 to form a neural network. According to the neural network of the present embodiment, in addition to the effects of the above-described embodiment, the scale of the neuron element 21 can be reduced, so that the neural network can be constructed in a large scale.

An embodiment of the invention described in claim 4 will be described with reference to FIG. In this embodiment, a latch 22 is added to the output line of the multiplier 14 in the configuration of the neuron element 18 shown in FIG. 4 to form a neuron element 23. The presence of the latch 22 enables pipeline processing, and the multiplication processing by the multiplier 14 and the addition processing by the adder 12 can be performed at the same time.

As in the case of the neuron element 11, a plurality of such neuron elements 23 are prepared and connected to each other in accordance with FIG. 3 to form a neural network. According to the neural network of the present embodiment, in addition to the effect of the above-described embodiment using the neuron element 18, the processing speed of the neuron element 21 can be improved and the flexibility of the neural network system can be increased. it can.

As shown in FIG. 7, a latch 22 may be provided on the output line of the multiplier 14 in the neuron element 21 shown in FIG. 5 to form the neuron element 24 (claim 3). It corresponds to the invention described in claim 4 with respect to the described invention).

As described above, the neuron elements 11 and 1 having various configurations are provided.
Although the embodiments of the neural network using 8, 21, 23, and 24 have been described, these neural networks can output various signals depending on parameters such as the coupling coefficient w _i and the threshold θ. Therefore, in applying such a neural network, an embodiment relating to a signal processing device for determining parameters such as a coupling coefficient w _i and a threshold value θ for outputting a signal suitable for the application will be shown below.

First, an embodiment of the invention described in claim 6 will be described with reference to FIG. In this embodiment, a genetic algorithm (Genetic Algorithm) is used to update the parameters. The signal processing device 25 of this embodiment is a neural network (NN) according to any of the embodiments described above.
W) 26 (that is, any one of the neuron elements 11, 18, 21, 23 or 24 is used as a neuron element) and a genetic algorithm computing means (GA) 27 which serves as a parameter updating means. There is. The neural network 26 receives a signal from the outside world through an input line 28 and outputs the signal to the outside world through an output line 29. The input side of the neural network 26 is connected to the input line 28 and the output line 29 of the genetic algorithm calculating means 27, and the output line 30 of the genetic algorithm calculating means 27 is connected to the neural network 26.
It is connected to the.

With such a configuration, the neural network 26 inputs a signal from the outside world through the input line 28,
The processed signal is output from the output line 29 to the outside world. At this time, the genetic algorithm computing means 27 inputs signals from these input lines 28 and output lines 29 and calculates and updates parameters such as the coupling coefficient w _i and the threshold value θ from these signals,
The updated parameters are loaded from output line 30 into neural network 26.

At this time, the genetic algorithm calculating means 27
The genetic algorithm calculation process performed by will be described. The coupling coefficient w _i and the threshold value θ are coded in the gene. That is, these values are used as the loci values. Generates M genes. Genes are sequentially loaded into the neural network 26, and an input line 28 and an output line 29 when each gene is used
Calculate the fitness from the above signal. M genes are selected and stored with a probability according to the fitness. However, the L highest fitness values are preferentially stored. Cross genes. Mutate a gene with probability p. The locus causing mutation is added to the original value by the stochastic mutation amount δ (δ is a real number). Return to. If the end conditions are met, the process ends.

According to this embodiment, the neural network 26 is provided with the genetic algorithm calculating means 27.
Since the parameters used by can be updated, the applicability as a neural network system can be enhanced.

As shown in FIG. 9, the connection of the output line 29 from the neural network 26 to the genetic algorithm calculating means 27 is omitted, and only the input signal to the neural network 26 is input to the genetic algorithm calculating means 27. It may be configured as the signal processing device 31 for converting to a signal (corresponding to the invention of claim 7). Also in this case, the same effect as in the case of FIG. 8 is obtained.

An embodiment of the invention described in claim 8 will be described with reference to FIG. In this embodiment, random search is used to update the parameters. The signal processing device 32 of the present embodiment is composed of the neural network (NNW) 16 according to any of the above-mentioned embodiments and the random search calculation means 33 serving as the parameter updating means. That is, the random search calculation means 33 is provided in place of the genetic algorithm calculation means 27 in FIG. 8, and the input side 28 and the output line 29 of the neural network 26 are connected to the input side of the random search calculation means 33. The output line 34 of the random search calculation means 33 is connected to the neural network 26.

With this configuration, the neural network 26 inputs a signal from the outside world through the input line 28,
The processed signal is output from the output line 29 to the outside world. At this time, the random search calculation means 33 receives these input lines 28,
The signals from the output line 29 are input, parameters such as the coupling coefficient w _i and the threshold value θ are calculated and updated from these signals, and the updated parameters are loaded from the output line 34 to the neural network 26.

At this time, the random search calculation process performed by the random search calculation means 33 will be described. The coupling coefficient w _i and the threshold value θ are set as one set. Generate M sets. The sets are sequentially loaded into the neural network 26, and the fitness is calculated from the signals on the input line 28 and the output line 29 when each set is used. M sets are selected and stored with a probability according to the fitness.
However, the L highest fitness values are preferentially stored. Mutate the set with probability p. The elements that cause mutation are added to the original value by the stochastic variation amount δ (δ is a real number). Return to. If the end conditions are met, the process ends.

According to this embodiment, since the random search operation means 33 is provided and the parameters used by the neural network 26 can be updated, the applicability as a neural network system can be enhanced. In particular, since the parameter updating means can be simplified as compared with the case of using the genetic algorithm computing means 27, the hardware implementation becomes easier.

As shown in FIG. 11, the connection of the output line 29 from the neural network 26 to the random search calculation means 33 is omitted and the neural network 26 is omitted.
It may be configured as the signal processing device 35 in which only the input signal with respect to is used as the input signal of the random search calculation means 33 (corresponding to the invention of claim 9). Also in this case, FIG.
The same effect as in the case of 0 can be obtained.

Further, these signal processing devices 25, 31,
Both 32 and 35 have one neural network 2
However, it is also possible to prepare a plurality of neural networks and use these neural networks in parallel with the genetic algorithm computing means 27 or the random search computing means 33 (claim). Which corresponds to the invention of item 10). According to this
The optimum value of the parameter can be searched at high speed, and the processing speed can be improved. In particular, when the input / output signals of a plurality of neural networks interact with each other, improved performance can be expected.

Next, starting with the above-mentioned neural network 16, the point is to show an example of using a neural network whose state dynamically changes in the following embodiments. First, an embodiment of the invention described in claims 11 and 12 will be described. In this embodiment, a neural network whose state dynamically changes, for example, the above-mentioned neural network 26 is used to construct an autonomous system with high autonomy. That is, the system configuration has a signal input means for inputting a signal to the neural network 26 and a signal output means for inputting a signal output from the neural network 26, and an output dynamically transiting to a certain input is generated. It is a thing.

An embodiment of the invention described in claims 13 to 16 will be described with reference to FIGS. 12 to 15. In this embodiment, a chaotic neural network as shown below is used instead of the neural network 26 in constructing an autonomous system. FIG. 12 shows a configuration example of one neuron element 40 which is an element constituting the chaotic neural network of this embodiment. First, an adder (adding means) 41 for adding a plurality of signals and a non-linear conversion section (non-linear conversion means) 4 for non-linearly converting the addition result
2 and an output unit (output means) for outputting the conversion result to the outside
43 are provided in order. A plurality of external input sections (external input means) 44, a first internal input section (first internal input means) 45, and a second internal input section (second internal input means) are provided on the input side of the adder 41. ) 46 and a memory (storage means) 47 storing a predetermined input signal are connected in parallel,
Each signal can be input to the adder 41. A chaotic neural network is constructed by connecting a plurality of such neuron elements 40 to each other.

Here, the external input section 44 multiplies an external input signal and a signal stored in the memory 48, specifically, a coupling coefficient by a multiplier 49, as shown in FIG. The multiplication result is input to the adder 41. The first internal input unit 45 multiplies an output signal 50 from the adder 41 and a signal stored in the memory 51, specifically, a predetermined coefficient by a multiplier 52, as shown in FIG. Then, this multiplication result is input to the adder 41. The second internal input unit 46 includes a nonlinear conversion unit 4 as shown in FIG.
The output signal 53 from 2 and a signal stored in the memory 54, specifically, a predetermined coefficient are multiplied by a multiplier 55, and the multiplication result is input to the adder 41. The output unit 43 is composed of a latch that holds the output value as the conversion result from the nonlinear conversion unit 42 for a certain period.

The non-linear conversion section 42 may convert the output signal from the adder 41 by using a table lookup, or may calculate the non-linear function by a DSP or a CPU. The above-mentioned adder 41 and multipliers 49, 52, 55 are also DSPs or CPs.
It may be realized by U, and the memory 4
7, 48, 51 and 54 may be registers that hold only one data. In addition, such a neuron element 40
May be realized partially or entirely by software.

According to such a chaotic neural network configuration, the output value of each neuron element 40 dynamically transits depending on the value stored in the memories 47, 48, 51 and 54 and the value of the external input signal. It exhibits a dynamic behavior, and a neural network having adaptability to a dynamic state can be constructed. Therefore, even when such a chaotic neural network is used, an autonomous system with high autonomy can be constructed as in the case of using the neural network 26. That is, a system configuration having a signal input means for inputting a signal to a chaotic neural network and a signal output means for inputting a signal output from the chaotic neural network is provided, and an output dynamically transiting to a certain input is generated. It is a thing.

Here, each neuron element 4 shown in FIG.
In the configuration of 0, the memory 47 may be deleted (corresponding to the invention of claim 14). Further, the memory 48 and the multiplier 49 may be substantially eliminated to form the first internal input section 44, and the output signal 50 from the adder 41 may be directly input to the adder 41 (claims). 1
Corresponding to the invention described in 5). Similarly, the memory 51 and the multiplier 52 may be substantially eliminated to form the second internal input unit 45, and the output signal 53 from the nonlinear conversion unit 42 may be directly input to the adder 41 (claim). Item 1
(Corresponding to the invention described in 6). According to these, the neural network can be constructed more easily.

Next, an embodiment of the invention described in claim 17 will be described with reference to FIGS. This embodiment uses a genetic algorithm to learn a chaotic neural network 56 having the neuron element 40 described in the above embodiment as a constituent element, that is, the memories 47, 48, 51 in each neuron element 40. The value of 54 is updated. First, the method of applying the genetic algorithm will be described. As shown in FIG. 16, suppose a gene whose factor is the value of the memory of each neuron element 40. Here, the memory 48 corresponds to the coupling coefficient, and there are as many external input units 44 as the neuron element 40.

In the genetic algorithm, the necessary data are
As described above, it is coded by considering it as one gene (individual). A plurality of such individuals are prepared to form a group. Then, the processing is performed according to the following procedures. Generating many genes (individuals). Each individual is loaded into the chaotic neural network 56, and the evaluation value Er is calculated from the output value for the training data. Here, the closer the evaluation value Er is to 0, the better the evaluation. Use the evaluation value, and "select" if it exceeds a certain standard value. Perform "crossover" between the remaining individuals. New individuals are generated by the number of selected individuals. It causes a mutation in an individual at a certain rate. The above steps are repeated, and the process ends when the optimum individual is found. The evaluation criteria and the mutation rate are arbitrarily determined empirically.

FIG. 17 shows an apparatus configuration for performing learning using such a GA, in which an evaluator (evaluation means) 57 is connected between the input and output of the chaotic neural network 56 and the chaotic neural network. GA calculator for 56 (genetic algorithm calculator)
58 is connected and comprised.

In such a configuration, the evaluator 57 inputs the test data as an input signal to the chaotic neural network 56 and inputs the output result obtained from the chaotic neural network 56. Therefore, the evaluator 57 uses the evaluation result Er from the output result of the chaotic neural network 56 and the given input signal to evaluate the value Er.
To calculate. The GA computing unit 58 writes and updates the values of the memories 47, 48, 51, 54 in the chaotic neural network 56 based on the calculated evaluation value Er.

According to this embodiment, since the chaotic neural network 56 can be learned, the flexibility becomes higher. Therefore, the flexibility of the autonomous system configured by using the chaotic neural network 56 is increased.

Next, an embodiment of the invention described in claim 18 will be described. This embodiment also relates to the learning method of the chaotic neural network 56 having the configuration as described in the above-mentioned embodiment, but the external input unit 4 is operated by a local learning rule such as Hebb's learning rule.
The value (coupling coefficient value) of the memory 48 in No. 4 is updated. That is, MEM _NEW = MEM _OLD + A × (IN-B, where MEM _NEW and MEM _OLD are values after updating the memories 48, 51, and 54, and IN is an input value multiplied by each memory value MEM _OLD. ) The value of the memory 48 is updated according to the following expression. Here, the values of A and B are empirically determined to be the memories 48, 51, and 56 in the chaotic neural network 56.
It is arbitrarily determined for each 54. In the learning method of this embodiment, the value in the memory 47 is fixed.

Also in the case of this embodiment, since the chaotic neural network 56 can be learned, it becomes more flexible. Therefore, the flexibility of the autonomous system configured by using the chaotic neural network 56 is increased.

Next, as a more concrete example of this autonomous system, an embodiment of the invention described in claims 19 and 21 to 26 will be described with reference to FIG. In this embodiment, the autonomous robot 59 is constructed by using a neural network whose state dynamically changes, for example, the chaotic neural network 56 described with reference to FIGS. That is, the signals from the sensors 60 of the autonomous robot 59 are the chaotic neural network 5
6 is applied to each drive system or display of the autonomous robot 59, here a display (drive display means) 61, as a control signal, and the output signal from each neuron element 40 in the output stage is driven or displayed. Is configured to let.

Here, the chaotic neural network 2
6 performs learning and self-organization in accordance with the above-described learning rule and evolves, and further, because the state transitions dynamically, it shows different movements even for signals from the same sensor 60. Therefore, a robot having adaptive autonomy can be realized.

In the construction of the autonomous robot 59 of this embodiment, instead of the chaotic neural network 56, the neural network 26 having the neuron element 11 (or 18, 21, 23 or 24) as a constituent element is used. It may be used (corresponding to the invention of claim 20).

FIG. 19 shows an embodiment of the invention according to claim 28.
22 to 22. In this embodiment, an autonomous traveling robot 62 using the signal processing device 25, 31, 32 or 35 as described above with reference to FIGS. 8 to 11 is configured as a mobile system. First, FIG. 19 is a schematic diagram showing the external configuration of the autonomous mobile robot 62.
The robot body 63 is configured to be movable by front wheels (steering wheels) 64 and rear wheels (driving wheels) 65. Further, on the robot body 63, a sensor (potential detection means) 66 such as an electrometer and an altimeter for detecting the potential (potential, altitude, etc.) of the work area is mounted.

FIG. 20 shows such an autonomous mobile robot 62.
FIG. 8 is a block diagram configuration of a controller system 67 including a neural network 26 (the neural network 26 itself has the configuration illustrated in FIG. 3) as shown in FIGS. 8 to 11 and a parameter updating means 27 or 33. (Substantially, the signal processing devices 25, 31, 3
2 or 35) is provided and is configured to take in the potential signal from the sensor 66. A decoder (signal conversion means) 68 is provided on the output side of the controller system 67 and is divided into two signals for the front wheels 64 and the rear wheels 65. The front wheel 64 is a signal showing the amount of movement in the left-right direction, the rear wheel 65 is a signal showing the amount of movement in the front-rear direction, and the decoder 68 converts the output from the controller system 67 into a steering angle and a drive amount. The front wheel 64 and the rear wheel 65 are actually controlled respectively.

The autonomous mobile robot 62 having such a configuration
Let us consider a case where is quickly moved to a region having the highest potential in the work region. Therefore, the sum of potential inputs in the work area is used as the fitness used for updating the parameters. FIG. 21 shows an example of the potential of the work area.

If the autonomous mobile robot 62 is left in such a work area, the system automatically moves to a place having the highest potential. The locus T of the autonomous mobile robot 62 as described above is, for example, as shown in FIG. 22, and the chaotic search (random search) is performed on the potential flat land A, and the hill climbing method is performed from the foot C of the potential mountain B. The target is searched (the steepest descent method), reaches the top D of the potential, and stays in the vicinity. That is, the movement trajectory T shown in FIG. 22 is a simulation for controlling the autonomous traveling robot 62 by the neural network 26 which interconnects the three neuron elements 11A to 11C as shown in FIG. 3, for example. This is due to the results of the generational robots. A genetic algorithm was used to update the parameters such as the coupling coefficient.

In other words, in a moving system having only a single sensor such as the sensor 66, it is very difficult with ordinary software to execute a task of moving to an area having the highest automatic potential. According to the embodiment, for example, only three neuron elements 11A
A non-equilibrium neural network of ~ 11C can realize a highly non-linear task.

In this embodiment, the autonomous traveling robot 62 using the signal processing device 25, 31, 32 or 35 provided with the parameter updating means 27 or 33 has been described, but it is constructed by using only the neural network 26. It may be the one (corresponding to the invention of claim 27).

FIG. 23 shows an embodiment of the invention according to claim 29.
Will be described. In the above embodiment, the autonomous traveling robot 6
Although the controller system 67 mounted in 2 is configured as an offline system including the parameter updating means 27 or 33, in the present embodiment, the controller system 69 using only the neural network 26 is used, and the parameter updating means 27 or 33 is a remote host. The computer has an online system in which signals are transmitted and received by the communication means 70. In such an online system, the individual used by the parameter updating means 27 or 33 may be assigned to a plurality of robots (autonomous traveling robot 62).

[0098]

According to the neural network of the first aspect of the present invention, a plurality of multi-input / single-output neuron elements in which the displacement amount of the internal activation value is determined according to the neuron output value are connected to each other. So steady state, chaos,
Since it becomes possible to show a dynamic state such as a limit cycle, a complicated system can be easily constructed with a simple network.

According to the neural network of the second aspect of the present invention, since the internal activation value storage means is provided, a plurality of signals can be input to the same signal line in time series, and the neural network according to the first aspect of the present invention is described. The expandability and processing speed of the neural network of the invention can be improved, and the flexibility of the system can be improved.

According to the neural network of the invention described in claim 3, the storage means for storing the threshold value for determining the displacement amount of the internal activation value is also used as the storage means for the coupling coefficient. The size of the neuron element in the neural network of the described invention can be reduced,
It is possible to promote the scale-up of the neural network.

According to the neural network of the invention described in claim 4, since the latch means is provided on the output line from the external input means to the adder means, pipeline processing becomes possible, the processing speed is improved, and the system is flexible. The property can be further improved.

According to the neural network of the fifth aspect of the present invention, the displacement amount of the internal activation value is calculated by (sum of products of input value of external input signal and coupling coefficient value) + (threshold value) − (internal activation value). And the neuron output value). Therefore, the neural network according to the present invention can be easily realized.

According to the signal processing device of the present invention as defined in claim 6 or 7, it is possible to realize the signal processing device according to claim 1, 2, 3, 4 or 5 as described above.
A parameter updating means by a genetic algorithm calculating means for inputting an input / output signal or an input signal of the neural network to the neural network of the described invention, calculating parameters such as a coupling coefficient and a threshold value of the neural network, and outputting to the neural network. Since it is possible to update the parameters of the neural network, the applicability of the system can be improved.

Also in the signal processing device of the invention of claim 8 or 9, the input / output signal or input of this neural network is applied to the neural network of the invention of claim 1, 2, 3, 4 or 5 as described above. Neural network coupling coefficient, the parameters such as threshold value are calculated by inputting the signal and output to this neural network, because the parameter updating means by the random search operation means is combined, it is possible to update the parameters of the neural network. The applicability of the system can be enhanced, and the parameter updating means can be simpler than the case of using the genetic algorithm computing means, and hardware implementation can be facilitated.

According to the signal processing device of the tenth aspect of the present invention, since a plurality of neural networks are provided, the processing speed of the signal processing device of the above-mentioned sixth, seventh, eighth or ninth aspect of the present invention is improved. It is possible to improve the performance, and if the input / output signals of a plurality of neural networks are allowed to interact with each other, the performance can be improved.

According to the autonomous system of the invention described in claims 11 to 18, since the autonomous system is configured by using the neural network whose state dynamically changes, it is possible to construct a highly autonomous system.

According to the autonomous robot of the invention described in claims 19 to 26, since the autonomous robot is configured by using the neural network whose state dynamically changes, it is possible to provide a highly autonomous robot.

According to the mobile system of the twenty-seventh to twenty-ninth aspects of the invention, the mobile system is constructed by using a neural network whose state dynamically changes or a signal processing device equipped with such a neural network.
With a very simple structure, it is possible to build a complicated mobile system.

[Brief description of drawings]

FIG. 1 is a block diagram of one neuron element showing an embodiment of the invention described in claims 1 and 5. FIG.

FIG. 2 is a schematic diagram showing an input / output relationship of one neuron element.

FIG. 3 is a schematic diagram showing a configuration example of a neural network.

FIG. 4 is a block diagram of one neuron element showing an embodiment of the invention described in claim 2;

FIG. 5 is a block diagram of one neuron element showing an embodiment of the invention according to claim 3;

FIG. 6 is a block diagram of one neuron element showing an embodiment of the invention described in claim 4;

FIG. 7 is a block diagram of one neuron element showing a modification thereof.

FIG. 8 is a block diagram of a signal processing device showing an embodiment of the invention described in claim 6;

FIG. 9 is a block diagram of a signal processing device showing an embodiment of the invention described in claim 7;

FIG. 10 is a block diagram of a signal processing device showing an embodiment of the invention described in claim 8;

FIG. 11 is a block diagram of a signal processing device showing an embodiment of the invention described in claim 9;

FIG. 12 is a block diagram of one neuron element showing an embodiment of the invention described in claims 13 to 16;

FIG. 13 is a block diagram showing the configuration of the external input unit.

FIG. 14 is a block diagram showing a configuration of a first internal input section thereof.

FIG. 15 is a block diagram showing a configuration of a second internal input section thereof.

FIG. 16 is an explanatory diagram of genes showing an example of the invention of claim 17;

FIG. 17 is a block diagram.

FIG. 18 is a schematic configuration diagram showing an embodiment of the invention described in claims 19 and 20.

FIG. 19 is a schematic configuration diagram of an autonomous traveling robot showing an embodiment of the invention as set forth in claim 28.

FIG. 20 is a block diagram thereof.

FIG. 21 is a characteristic diagram showing a potential example of a work area.

FIG. 22 is a plan view showing a movement trajectory in a work area.

FIG. 23 is a block diagram showing an embodiment of the invention as set forth in claim 29.

FIG. 24 is a block diagram of a chaotic neuron showing a conventional example.

[Explanation of symbols]

11 neuron element 12 adding means 13 non-linear converting means 14 external input means 15 internal input means 17 storage means 18 neuron element 19 storage means 21 neuron element 22 latch 23, 24 neuron element 25 signal processing device 26 neural network 27 genetic algorithm computing means Parameter updating means 31 and 32 Signal processing device 33 Parameter updating means by random search computing means 35 Signal processing device 40 Neuron element 41 Adder means 42 Non-linear conversion means 43 Output means 44 External input means 45 First internal input means 46 Second internal Input means 47 Storage means 56 Chaotic neural network 57 Evaluation means 58 Genetic algorithm calculation means 59 Autonomous robot 60 Sensor 61 Drive display means 62 Moving system 63 Transfer Body 64 steering wheel 65 the drive wheel 66 potential detection means 67 signal processor 68 the signal converting means 69 signal processing device 70 the communication means

Claims (29)

[Claims] 1. An addition means for adding a plurality of signals, a non-linear conversion means for non-linearly converting an internal activation value resulting from the addition of the addition means, and outputting the conversion result to the outside as a neuron output value, and an external input signal. And an external input means for multiplying the coupling coefficient and outputting the multiplication result to the adding means, and an internal input means for performing arithmetic processing on the internal activation value and the neuron output value and outputting to the adding means. A neural network comprising a plurality of multi-input / one-output neuron elements whose displacements of the internal activation value are determined according to the neuron output value, which are connected to each other. 2. The neural network according to claim 1, further comprising storage means connected to the addition means for storing the internal activation value. 3. The neural network according to claim 2, wherein the storage means for storing the threshold value for determining the displacement amount of the internal activation value is also used as the storage means for the coupling coefficient. 4. A latch means is provided on the output line from the external input means to the adder means.
Neural network described. 5. The displacement amount of the internal active value is calculated by (sum of products of input value of external input signal and coupling coefficient value) +
5. The neural network according to claim 1, 2, 3, or 4, wherein (threshold value)-(product of internal activation value and neuron output value). 6. The neural network according to claim 1, 2, 3, 4 or 5, and the input / output signals of the neural network as input to calculate parameters such as a coupling coefficient and a threshold of the neural network, and the neural network. A signal processing device comprising: a parameter updating means by a genetic algorithm calculating means for outputting to. 7. The neural network according to claim 1, 2, 3, 4 or 5, and parameters such as a coupling coefficient and a threshold value of the neural network are calculated by using an input signal for the neural network as an input to the neural network. A signal processing device comprising: a parameter updating unit for outputting a genetic algorithm calculating unit. 8. The neural network according to claim 1, 2, 3, 4 or 5, and the input / output signals of the neural network as inputs to calculate parameters such as a coupling coefficient and a threshold of the neural network, and the neural network. A signal processing device comprising: a parameter updating unit that outputs a random search calculation unit that outputs to the. 9. The neural network according to claim 1, 2, 3, 4 or 5, and the input signal to the neural network is used as an input to calculate parameters such as a coupling coefficient and a threshold value of the neural network, A signal processing device, comprising: a parameter updating unit for outputting a random search calculation unit. 10. The signal processing device according to claim 6, further comprising a plurality of neural networks. 11. A neural network that dynamically changes states, a signal input means for inputting a signal to the neural network, and a signal output means for inputting a signal output from the neural network. Autonomous system to do. 12. The autonomous system according to claim 11, wherein the neural network is the neural network according to claim 1, 2, 3, 4 or 5. 13. A neural network, adding means for adding a plurality of signals, non-linear converting means for performing non-linear conversion of the addition result of the adding means, and output means for outputting the conversion result by the non-linear converting means to the outside. An external input means for multiplying a plurality of external input signals by a coupling coefficient and outputting the multiplication result to the adding means; and a second means for multiplying the addition result of the adding means by a coefficient and outputting the multiplication result to the adding means. A chaos in which a plurality of neuron elements having one internal input means and a second internal input means for multiplying the conversion result of the non-linear conversion means by a coefficient and outputting the multiplication result to the addition means are interconnected. The autonomous system according to claim 11, which is a neural network. 14. The autonomous system according to claim 13, wherein each neuron element in the chaotic neural network has a storage means for storing a signal to be input to the addition means. 15. The autonomous system according to claim 13, wherein the addition result of the adding means is directly output to the adding means as the first internal input means. 16. The autonomous system according to claim 13, 14 or 15, wherein the conversion result of the non-linear conversion means is directly output to the addition means as the second internal input means. 17. A chaotic neural network comprising:
Evaluation means for generating an input signal and inputting an output result to generate an evaluation value from the input signal and the output result, and a coefficient inside the neuron element is read out by using the evaluation value by the evaluation means and the coefficient is read out. 17. The autonomous system according to claim 13, 14, 15 or 16, further comprising: a genetic algorithm calculation means that updates the coefficient according to the genetic algorithm and writes the updated coefficient to each neuron element. 18. An external input means for updating a coupling coefficient by using an input signal.
The autonomous system according to 4, 15 or 16. 19. A neural network that dynamically changes its state, a sensor that generates a signal to be input to this neural network, and a drive display means that receives or outputs a signal output from the neural network as a control signal. An autonomous robot having: 20. The autonomous robot according to claim 19, wherein the neural network is the neural network according to claim 1, 2, 3, 4 or 5. 21. An addition means for adding a plurality of signals to a neural network, a non-linear conversion means for performing non-linear conversion of the addition result of the addition means, and an output means for outputting the conversion result of the non-linear conversion means to the outside. An external input means for multiplying a plurality of external input signals by a coupling coefficient and outputting the multiplication result to the adding means; and a second means for multiplying the addition result of the adding means by a coefficient and outputting the multiplication result to the adding means. A chaos in which a plurality of neuron elements having one internal input means and a second internal input means for multiplying the conversion result of the non-linear conversion means by a coefficient and outputting the multiplication result to the addition means are interconnected. 20. The autonomous robot according to claim 19, which is a neural network. 22. The autonomous robot according to claim 21, wherein each neuron element in the chaotic neural network has a storage means for storing a signal to be input to the addition means. 23. The autonomous robot according to claim 21, wherein the addition result of the adding means is directly output to the adding means as the first internal input means. 24. The autonomous robot according to claim 21, 22 or 23, wherein the second internal input means directly outputs the conversion result of the non-linear conversion means to the addition means. 25. In a chaotic neural network,
Evaluation means for generating an input signal and inputting an output result to generate an evaluation value from the input signal and the output result, and a coefficient inside the neuron element is read out by using the evaluation value by the evaluation means and the coefficient is read out. 25. The autonomous robot according to claim 21, 22, 23 or 24, further comprising: a genetic algorithm calculation means that updates the coefficient according to a genetic algorithm and writes the updated coefficient in each neuron element. 26. An external input means for updating a coupling coefficient by using an input signal.
The autonomous robot according to 2, 23 or 24. 27. A potential detecting means for detecting a potential of a work area, a neural network according to claim 1, 2, 3, 4 or 5, which receives a detection output from this potential detecting means, and a neural network from this neural network. A moving system comprising: a signal converting means for converting and outputting an output signal; and steered wheels and driving wheels to which output signals from the signal converting means are given as control signals. 28. The potential detecting means for detecting the potential of the work area, and the detection output by the potential detecting means as inputs.
A signal processing device as described above, a signal converting device for converting and outputting an output signal from the signal processing device, and steering wheels and driving wheels to which the output signal from the signal converting device is given as a control signal. A characteristic moving system. 29. The mobile system according to claim 27, wherein the parameter updating means is provided outside the mobile body and is connected by communication means.