Memcapacitor classical conditioned reflex generalization differentiation neural network circuit

Abstract

The invention relates to the technical field of generalization differentiation neural networks, in particular to a memcapacitor classic conditioned reflex generalization differentiation neural network circuit, which comprises an acid signal circuit, a memory capacitor classic conditioned reflex generalization differentiation neural network circuit and a memory capacitor classic conditioned reflex generalization differentiation neural network circuit, the grabbing signal circuit is used for adjusting an input grabbing action signal; the flexible signal circuit is used for adjusting an input flexible action signal; the salivation circuit is used for detecting whether leakage current is generated or not and judging whether training is successful or not; by adjusting the input states of the three signals and the on-off of the switch, the generalization and differentiation function in the classical conditioned reflex of the animal is trained and realized. The invention solves the problem of lack of a memory capacitor neural network circuit for regulating and controlling the capacitance value of the memory capacitor in the existing design.

Description

Technical Field

[0001] The present invention relates to the technical field of generalized differentiation neural network, and in particular to a memory capacity classical conditioning reflex generalized differentiation neural network circuit. Background Art

[0002] A memcapacitor is a capacitor that can store charge and maintain the stored charge when the power is off. The working principle of a memcapacitor is similar to that of a memristor, both of which are based on the electrical properties of the material.

[0003] Memcontainers have broad application prospects and can be used in electronic storage, sensors, energy harvesters and other fields. In terms of electronic storage, memcontainers can achieve higher storage density and lower energy consumption, while also improving the reliability and stability of storage. In the future, with the continuous development and improvement of memcontainer technology, it will be applied in more fields, bringing more convenience and benefits to mankind.

[0004] Many people have proposed the classical conditioning neural network circuit of memristor. For example, the literature Amemristor-based circuit design for generalization and differentiation on Pavlov associative memory also involves the generalization and differentiation neural network circuit. However, the literature has the following problems in using memristor for training:

[0005] 1. Lack of design of memcapacitor neural network circuit by adjusting the capacitance value of memcapacitor;

[0006] 2. The input signal is a pulse signal, which has the problem of high static power consumption;

[0007] 3. Using operational circuits to determine whether there is leakage current and then reflect the learning effect is not simple and intuitive enough. Summary of the invention

[0008] In view of the deficiencies of the existing circuits, the present invention solves the problem that the existing designs lack a memcapacitor neural network circuit that can adjust the capacitance value of the memcapacitor.

[0009] The technical solution adopted by the present invention is: the memory capacity classical conditioning reflex generalization differentiation neural network circuit includes: an acid signal circuit, a grasping signal circuit, a scratching signal circuit and a salivation circuit; wherein,

[0010] The acid signal circuit is used for adjusting the input acid stimulation signal;

[0011] The grasping signal circuit is used to adjust the input grasping action signal;

[0012] The bending signal circuit is used to adjust the input bending action signal;

[0013] The salivation circuit is used to detect whether leakage current is generated and determine whether the training is successful;

[0014] By adjusting the input status of the three signals and the on-off of the switch, training can achieve the function of generalization and differentiation in the animal's classical conditioned reflex.

[0015] As a preferred embodiment of the present invention, the function of generalization differentiation in the classical conditioning of the stimulating action includes: switch S1 is closed, switch S2 is opened; V acid Input, V rub and V scratch No input causes the gate voltage of NMOS tube N1 to exceed the threshold voltage, generating leakage current, which corresponds to the salivation of animals when they receive acidic stimulation in the generalization differentiation of classical conditioning;

[0016] V rub Input, V acid and V scratch No input makes the gate voltage of NMOS tube N1 less than the threshold voltage, and no leakage current appears, which corresponds to the part in the generalization differentiation of classical conditioning reflex where the animal does not salivate after receiving the action of being caught;

[0017] V scratch Input, V acid and V rub No input makes the gate voltage of NMOS tube N1 less than the threshold voltage, and no leakage current occurs, which corresponds to the part in the generalization differentiation of classical conditioning where the animal does not salivate when being tickled.

[0018] As a preferred embodiment of the present invention, the function of generalization differentiation in the classical conditioning of the stimulating action also includes:

[0019] Switch S1 is open, switch S2 is closed; V acid 、V rub Input, V scratch No input causes the gate voltage of NMOS tube N1 to exceed the threshold voltage, generating leakage current, which corresponds to the salivation part when the animal receives acidic stimulation and grasping action at the same time in the generalization of classical conditioning;

[0020] V rub Input, V acid and V scratch No input causes the gate voltage of NMOS tube N1 to exceed the threshold voltage, generating leakage current, which corresponds to the animal establishing a connection between acidic stimulation and grasping action in the generalization of classical conditioning, and the part that will also produce salivation when being grasped;

[0021] V scratch Input, V acidand V rub No input causes the gate voltage of the NMOS tube N1 to exceed the threshold voltage, generating leakage current. This corresponds to the fact that in the generalization of classical conditioning, when the animal is subjected to a scratching action similar to scratching, it will also produce salivation, indicating that generalization is successful.

[0022] As a preferred embodiment of the present invention, the acid signal circuit includes: an operational amplifier U1, resistors R1, R8, a switch S1, and a capacitor C1. The same-direction input terminal of U1 is connected to the acid signal, the reverse input terminal of U1 is connected to the output terminal of U1, the left ends of R1 and R2, the right end of R1 is connected to the input terminal of S1, the output terminal of S1 is respectively connected to the left end of C1 and the left end of R8, and the right end of R8 is connected to the right end of R9.

[0023] As a preferred embodiment of the present invention, the grasping signal circuit includes: the bending signal circuit includes: resistors R2 to R7, R9, R 10 , R 19 , operational amplifiers U2 to U5 and memory capacitor MC1, the right end of R2 is respectively connected to the right end of R3, the inverting input of U2 and the left end of R4, the left end of R3 is connected to the captured signal, the right end of R4 is respectively connected to the output of U2 and the left end of R5, the right end of R5 is respectively connected to the inverting input of U3 and the left end of R6, the right end of R6 is respectively connected to the output of U3 and the left end of R7, the output of U3 is also connected to the input of S2, the right end of R7 is respectively connected to the left end of MC1 and the same-direction input of U4, the inverting input of U4 is respectively connected to the output of U4 and the left end of R9, the right end of R9 is respectively connected to R 10 The left end of U5 is connected to the reverse input terminal, R 10 The right end is connected to the output end of U5, R 19 The left end of the connection.

[0024] As a preferred embodiment of the present invention, the signal circuit includes: a resistor R 11 To R 18 , R 20 , operational amplifiers U6 to U9, memcapacitor MC2 and switch S2, R 11 The right end of R 12 The right end of U6, the inverting input terminal and R 13 The left end of the connection, R 12 The left end is connected to the torsion signal, R 13 The right end is connected to the output end of U6 and R 14 The left end of the connection, R 14 The right end is connected to the reverse input terminal of U7 and R 15 The left end of the connection, R 15 The right end is connected to the output end of U7 and R 16 The left end of the connection, R16 The right end of is connected to the left end of MC2 and the same-direction input end of U8, and the reverse input end of U8 is connected to the output end of U8 and R 17 The left end of the connection, R 17 The right end of R 18 The left end of the U9 is connected to the reverse input terminal, R 18 The right end is connected to the output end of U9, R 20 The left end of the connection, R 11 The left end of is connected to the output end of S2.

[0025] As a preferred embodiment of the present invention, the salivation circuit includes: an operational amplifier U 10 , resistor R 21 , R 22 and NMOS tube N1, R 21 The left end of R 19 , R 20 The common right end of U 10 The inverting input terminal is connected to R 21 The right end of 10 The output terminal of N1 is connected to the gate of N1, and the drain of N1 is connected to R 22 The lower end of the connection, R 22 The upper end is connected to a DC voltage.

[0026] As a preferred implementation manner of the present invention, the model of the operational amplifier is AD8538.

[0027] As a preferred implementation of the present invention, the model of the NMOS tube is M2SK530.

[0028] Beneficial effects of the present invention:

[0029] 1. By designing acidic signal V acid , captured signal V rub , the signal V scratch To regulate the capacitance value of the memory container and realize the function of generalization and differentiation in the classical conditioning of organisms, it has great significance in the field of memory capacity neural network circuits. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a schematic diagram of a circuit for generalization and differentiation of the memory capacity classical conditioning reflex of the present invention;

[0031] Figure 2 It is a simulation diagram of the initial test circuit for the generalization differentiation of the memory capacity classical conditioning reflex of the present invention;

[0032] Figure 3 It is a circuit simulation diagram of the generalization stage in the generalization differentiation of the memory classical conditioning reflex of the present invention;

[0033] Figure 4 This is a circuit simulation of the differentiation stage in the generalization differentiation of the memory classical conditioning reflex of the present invention. DETAILED DESCRIPTION

[0034] The present invention is further described below in conjunction with the accompanying drawings and embodiments. This figure is a simplified schematic diagram, which only illustrates the basic structure of the present invention in a schematic manner, and therefore it only shows the components related to the present invention.

[0035] like Figure 1 As shown, the memory capacity classical conditioning reflex generalization differentiation neural network circuit includes:

[0036] Input signal acid signal V acid , captured signal V rub , the signal V scratch The salivation effect in the animal's classical conditioned reflex is reflected by whether the NMOS tube is turned on and whether there is leakage current passing through; V acid One branch is connected to capacitor C1 through a voltage follower, and the other branch is connected to the captured signal V rub Connected to the memory capacitor MC1 through an adder and a voltage follower; V acid and catch the signal V rub After summing with V scratch It is connected to the memcapacitor MC2 through an adder and a voltage follower; the voltage across the capacitor C1 and the voltage signal across the memcapacitor MC1 are summed together through an adder, and then summed again with the voltage signal across the memcapacitor MC2, and output to the gate of the NMOS tube; the capacitance values ​​of MC1 and MC2 will change accordingly with different input signals, which will lead to changes in the voltage across the memcapacitor, and ultimately determine whether there is leakage current in the NMOS tube. Whether there is leakage current is observed to determine whether the circuit conforms to the process of generalization and differentiation of animal classical conditioned reflexes.

[0037] The components used in the circuit are resistors R1, R7, and R 16 , the resistance R is 10kΩ 22 , the resistance value is 1kΩ resistor R2, R3, R4, R5, R6, R8, R9, R 10 , R 11 , R 12 , R 13 , R 14 , R 15 , R 17 , R 18 , R 19 , R 20 , R 21 , operational amplifiers U1, U2, U3, U4, U5, U6, U7, U8, U9, U 10, capacitor C1, memristors MC1 and MC2, switches S1 and S2, N-type MOS tube N1 with a threshold voltage of 2.1V; acid signal V acid The input of is connected to the positive input terminal of the operational amplifier U1, the inverting input terminal of the operational amplifier U1 is connected to the output terminal, the left end of the resistor R1 is connected to the output terminal of the operational amplifier U1, and the right end is connected to the input of the switch S1, the left end of the capacitor C1 is connected to the output terminal of the switch S1, and the right end is grounded; the positive input terminal of the switch S1 is connected to the control signal V con1 , the negative input terminal is grounded.

[0038] Catch the signal V rub The input is connected to the left end of the resistor R3, the right end of the resistor R3 is connected to the inverting input of the operational amplifier U2, the left end of the resistor R2 is connected to the output of the operational amplifier U1, and the right end is connected to the inverting input of the operational amplifier U2, the left end of the resistor R4 is connected to the inverting input of the operational amplifier U2, and the right end is connected to the output, and the positive input of the operational amplifier U2 is grounded; the left end of the resistor R5 is connected to the output of the operational amplifier U2, and the right end is connected to the inverting input of the operational amplifier U3, the left end of the resistor R6 is connected to the inverting input of the operational amplifier U3, and the right end is connected to the output of the operational amplifier U3, and the positive input of the operational amplifier U3 is grounded, the left end of the resistor R7 is connected to the output of the operational amplifier U3, and the right end is connected to the left end of the memory capacitor MC1, and the right end of the memory capacitor MC1 is grounded; the positive input of the switch S2 is connected to the control signal Vcon2, and the negative input is grounded; the output of the operational amplifier U3 is connected to the input of the switch S2, and the output of the switch S2 is connected to the resistor R 11 The left end is connected to the resistor R 11 The right end is connected to the inverting input terminal of the operational amplifier U6.

[0039] The signal V scratch Input terminal and resistor R 12 The left end is connected to the resistor R 12 The right end is connected to the inverting input of operational amplifier U6, and the resistor R 13 The left end is connected to the reverse input terminal of the operational amplifier U6, the right end is connected to the output terminal of the operational amplifier U6, the positive input terminal of the operational amplifier U6 is grounded, and the resistor R 14 The left end is connected to the output of the operational amplifier U6, and the right end is connected to the inverting input of the operational amplifier U7. 15 The left end is connected to the inverting input terminal of the operational amplifier U7, and the right end is connected to the output terminal of the operational amplifier U7. 16The left end is connected to the output end of the operational amplifier U7, and the right end is connected to the left end of the memcapacitor MC2, and the right end of the memcapacitor MC2 is grounded; the positive input end of the operational amplifier U4 is connected to the right end of the resistor R7, and the output end of the operational amplifier U4 is connected to the inverting input end. The left end of the resistor R9 is connected to the output end of the operational amplifier U4, and the right end is connected to the inverting input end of the operational amplifier U5. The left end of the resistor R8 is connected to the left end of the capacitor C1, and the right end of the resistor R8 is connected to the inverting input end of the operational amplifier U5. 10 The left end is connected to the inverting input of the operational amplifier U5, the right end is connected to the output of the operational amplifier U5, the positive input of the operational amplifier U8 is connected to the left end of the memory capacitor MC2, the output of the operational amplifier U8 is connected to the reverse input, and the resistor R 17 The left end is connected to the output of the operational amplifier U8, and the right end is connected to the inverting input of the operational amplifier U9. 18 The left end is connected to the inverting input terminal of the operational amplifier U9, the right end is connected to the output terminal of the operational amplifier U9, the positive input terminal of the operational amplifier U9 is grounded, and the resistor R 19 The left end is connected to the output of the operational amplifier U5, and the right end is connected to the output of the operational amplifier U 10 The inverting input terminal is connected to the resistor R 21 The left end of the operational amplifier U 10 The inverting input terminal of the operational amplifier U 10 The positive input terminal of the operational amplifier U 10 The output end of is connected to the gate end of NMOS tube N1, and the drain end of NMOS tube N1 is connected to the resistor R 22 The lower end of the resistor R 22 The upper end of is connected to a 3.5V DC bias, and the source end of the NMOS tube N1 is grounded.

[0040] The model of the operational amplifier is AD8538, and the model of the NMOS tube is M2SK530.

[0041] Circuit principle:

[0042] Control signal V con1 、V con2 The input signal includes a sine acid signal V with an amplitude of 2.2V and a frequency of 3HZ. acid ; The captured signal V is 3V with a frequency of 3HZ rub ; The amplitude is 2.9V, the frequency is 3HZ scratch ; V scratch 、V rub All are sinusoidal signals; V acidThe input is connected to switch S1 through a branch after the voltage follower U1. When switch S1 is closed, the input is input to the common capacitor C1. The other branch is connected to the captured signal V scratch They are connected to adder U2 and the output voltage V U3 Input to MC1, V U3 The expression is:

[0043] V U3 =V acid +V rub (1)

[0044] The capacitance of MC1 at voltage V U3 The capacitive reactance of MC1 is X MC1 Change, memory capacity resistance X MC1 The expression is:

[0045]

[0046] The change of the memcapacitive reactance makes the voltage V divided by the memcapacitor MC1 and the resistor R7. MC1 V MC1 The expression is:

[0047]

[0048] The voltage across the memory capacitor MC1 passes through the voltage follower U4 and the voltage across the capacitor C1 V C1 They enter adder U5 together, and the output voltage of adder U5 is V U5 , V U5 The expression is:

[0049] V U5 =V MC1 +V C1 (4)

[0050] The output voltage V after passing through the inverter U3 U3 When switch S2 is closed, V scratch The voltage V U7 Input to MC2, V U7 The expression is:

[0051] V U7 =V acid +V rub +V scratch (5)

[0052] The capacitance of MC2 at voltage V U7The capacitive reactance of MC2 is X MC2 Change, memory capacity resistance X MC2 The expression is:

[0053]

[0054] The change of memcapacitive reactance makes memcapacitor MC2 and resistor R 16 After voltage division, the voltage V divided across the capacitor is MC2 V MC2 The expression is:

[0055]

[0056] Voltage V MC2 After the voltage follower U8 and the inverter U9 and the voltage V U5 Input to adder U 10 , U 10 The output voltage is V U10 , V U10 The expression is:

[0057] V U10 =V U5 +V MC2 (8)

[0058] Voltage V U10 The output to the gate terminal of the NMOS tube N1 with a cut-off voltage of 2.1V determines the conduction and cut-off of the NMOS tube N1, so it can be observed whether there is a leakage current I D The emergence of the classical conditioning reflex can be used to judge the various stages of generalization and differentiation of the animal's classical conditioning reflex.

[0059] As attached Figure 2 As shown, during 0-15s, switch S1 is closed and switch S2 is open; during 0-5s, V acid is the input circuit, at this time V rub and V scratch Not input, V acid The voltage of capacitor C1 is input through voltage follower U1. C1 Through adder U5 and adder U 10 The output to the gate of NMOS tube N1 exceeds the threshold voltage of 2.1V, so leakage current I D , which corresponds to the part of the classical conditioning reflex generalization differentiation in which the animal salivates when it receives an acidic stimulus;

[0060] At 5-10s, V rub Input, at this time V acid and V scratch Not input, V rubThe voltage is input into the memcapacitor MC1 through the adder U2 and the voltage follower U3, which is lower than the negative threshold voltage -1.8V of the memcapacitor MC1 but not higher than the positive threshold voltage 3V. The memcapacitor value rises slightly. After the memcapacitor MC1 is divided by the resistor R7, the voltage V MC1 The voltage across the memory capacitor MC1 is V MC1 After passing through voltage follower U4 and adder U5, it enters adder U 10 , this voltage is less than 2.1V, and there is no leakage current I when it is output to the gate of NMOS tube N1 D The appearance corresponds to the part of the generalization differentiation of classical conditioning in which the animal does not salivate when it receives the grasping action;

[0061] At 10-15s, V scratch Input circuit, V acid and V rub Not input, V scratch After passing through adder U6 and inverter U7, it is input into memcapacitor MC2, which is lower than the negative threshold voltage -1.8V of memcapacitor MC2 but not higher than the positive threshold voltage 3V. The memcapacitor value rises slightly, and memcapacitor MC2 and resistor R 16 After voltage division, the voltage across the memory capacitor is V MC2 After passing through voltage follower U8 and inverter U9, it enters adder U 10 , this voltage is less than 2.1V, and there is no leakage current I when it is output to the gate of NMOS tube N1 D The appearance corresponds to the part of the generalization differentiation of classical conditioning in which the animal does not salivate when it is tickled.

[0062] As attached Figure 3 As shown, during 0-16s, switch S1 is open and switch S2 is closed; during 0-10s, V acid and V rub input circuit, while V scratch Not input, V acid Through the voltage follower U1 and V rub The sum is then added together through adder U2 to form a sinusoidal signal V with an amplitude of 5.2V and a frequency of 3HZ. U3 The signal is then output to the memcapacitor MC1 through a branch after the inverter U3. The signal voltage exceeds the positive threshold voltage 3V and negative threshold voltage -1.8V of the memcapacitor MC1. The memcapacitor value decreases faster than it increases. Therefore, the overall trend of the memcapacitor value decreases. The voltage V across the memcapacitor MC1 The voltage V formed by the voltage follower U4 and the adder U5 U5 Enter adder U 10 The other branch after the inverter U3 converts the sine signal V with an amplitude of 5.2V and a frequency of 3HZ U3The output is sent to adder U6, and then to memcapacitor MC2 through inverter U7. The signal voltage exceeds the positive threshold voltage 3V and negative threshold voltage -1.8V of memcapacitor MC2. The memcapacitor value decreases faster than it increases, so the overall trend of memcapacitor value decreases. Memcapacitor MC2 and resistor R 16 After voltage division, the voltage across the memory capacitor is V MC2 After passing through voltage follower U8 and inverter U9, it enters adder U 10 , and the voltage V U5 Input into adder U 10 , adder U 10 The output voltage exceeds 2.1V, so the leakage current I D The appearance corresponds to the part of salivation in the generalization of classical conditioning when the animal receives the acidic stimulus and the grasping action at the same time;

[0063] At 10-13s, V rub Input, V acid and V scratch Not input, V rub The voltage is input into the memory capacitor MC1 through the adder U2 and the voltage follower U3. After the voltage is divided by the memory capacitor MC1 and the resistor R7, the voltage V MC1 Exceeding the threshold voltage of NMOS tube N1 by 2.1V, it passes through voltage follower U4, adder U5 and adder U 10 When the output is sent to the gate of NMOS tube N1, the leakage current I D The appearance corresponds to the part in the generalization of classical conditioning in which the animal simultaneously associates the acidic stimulus with the action of grasping, and the part in which the grasping action also causes salivation;

[0064] At 13-16s, V scratch Input circuit, at this time V acid and V rub Not input, V scratch After passing through adder U6 and inverter U7, the input is fed into memory capacitor MC2. Memory capacitor MC2 and resistor R 16 After voltage division, the voltage across the memory capacitor is V MC2 When the voltage exceeds 2.1V, it passes through the voltage follower U8, inverter U9 and adder U 10 , when output to the gate of NMOS tube N1, the leakage current I D The appearance of this phenomenon corresponds to the generalization of classical conditioning, in which animals will salivate when they are subjected to a scratching action similar to that of scratching, indicating that generalization is successful.

[0065] As attached Figure 4 As shown, 0-32s ( Figure 4 yes Figure 3At this time, 0-32s is equivalent to 17-59s), switch S1 is disconnected, switch S2 is disconnected; at 0-3s, V acid and V rub input circuit, while V scratch Not input, V acid Through the voltage follower U1 and V rub The sum is then added together through adder U2 to form a sinusoidal signal V with an amplitude of 5.2V and a frequency of 3HZ. U3 The signal is then output to the memcapacitor MC1 through a branch after the inverter U3. The signal voltage exceeds the positive threshold voltage 3V and negative threshold voltage -1.8V of the memcapacitor MC1. The memcapacitor value decreases faster than it increases, so the memcapacitor value of MC1 continues to decrease, and the voltage V across the memcapacitor MC1 When the voltage exceeds 2.1V, the voltage V is formed after the voltage follower U4 and the adder U5. U5 Enter adder U 10 Then it is output to the gate of NMOS tube N1, and leakage current I D This corresponds to the differentiation of classical conditioning, in which the animal is simultaneously exposed to acidic stimulation and grasping action, and the connection between the two is deepened.

[0066] At 3-8s, V scratch Input circuit, V acid and V rub Not input, V scratch After passing through adder U6 and inverter U7, the voltage is input into memcapacitor MC2. The voltage is lower than the negative threshold voltage -1.8V of memcapacitor MC2 but does not exceed the positive threshold voltage 3V. Therefore, the memcapacitor value increases. The memcapacitor MC2 and the resistor R 16 After voltage division, the voltage across the memory capacitor is V MC2 When the voltage exceeds 2.1V, it passes through the voltage follower U8, inverter U9 and adder U 10 , when output to the gate of NMOS tube N1, the leakage current I D This corresponds to the action of the animal being tickled alone in the differentiation of classical conditioning. Although salivation still occurs when tickled, the connection between acidic stimulation and tickling has gradually weakened.

[0067] Repeat the operation of 0-3s at 8-11s, repeat the operation of 3-8s at 11-16s, repeat the operation of 0-3s at 16-19s, repeat the operation of 3-8s at 19-24s, and repeat the operation of 0-3s at 24-27s; repeat the operation of 3-8s at 27-30s. In the previous repeated operations, the capacitance of the memcapacitor MC2 has been slowly increasing; at 28s, the memcapacitor MC2 increases by 86nF. At this time, the memcapacitor MC2 and the resistor R 16 After voltage division, the voltage across the memory capacitor is VMC2 The voltage is lower than 2.1V, so it passes through the voltage follower U8, inverter U9 and adder U 10 , there is no leakage current I when output to the gate of NMOS tube N1 D This corresponds to the differentiation of classical conditioning, where the animal will gradually weaken the connection between the two when it is tickled several times without acidic stimulation, and will eventually stop salivating when tickled.

[0068] At 30-31 seconds, the input condition at 0-3 seconds is repeated, and there is a leakage current I at the gate end of the NMOS tube N1. D This corresponds to the classical conditioning differentiation in which after the animal strengthens the connection between acidic stimulation and grasping action, it will still salivate when given the grasping action alone, indicating that the animal has distinguished between grasping and scratching actions and the differentiation is successful.

[0069] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A neural network circuit for memory capacity classical conditioning reflex generalization differentiation, characterized in that: include: The acid signal circuit is used for adjusting the input acid stimulation signal; The grasping signal circuit is used to adjust the input grasping action signal; The bending signal circuit is used to adjust the input bending action signal; The salivation circuit is used to detect whether leakage current is generated and determine whether the training is successful; By adjusting the input status of the three signals and the on-off of the switch, training can achieve the function of generalization and differentiation in the animal's classical conditioned reflex.

2. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 1 is characterized in that: The functions of generalization differentiation in the classical conditioning of the action include: switch S1 is closed, switch S2 is opened; V acid Input, V rub and V scratch No input causes the gate voltage of NMOS tube N1 to exceed the threshold voltage, generating leakage current, which corresponds to the salivation of animals when they receive acidic stimulation in the generalization differentiation of classical conditioning; V rub Input, V acid and V scratch No input makes the gate voltage of NMOS tube N1 less than the threshold voltage, and no leakage current appears, which corresponds to the part in the generalization differentiation of classical conditioning reflex where the animal does not salivate after receiving the action of being caught; V scratch Input, V acid and V rub No input makes the gate voltage of NMOS tube N1 less than the threshold voltage, and no leakage current occurs, which corresponds to the part in the generalization differentiation of classical conditioning where the animal does not salivate when being tickled.

3. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 2 is characterized in that: The functions of generalization differentiation in classical conditioning of action also include: Switch S1 is open, switch S2 is closed; V acid 、V rub Input, V scratch No input causes the gate voltage of NMOS tube N1 to exceed the threshold voltage, generating leakage current, which corresponds to the salivation part when the animal receives acidic stimulation and grasping action at the same time in the generalization of classical conditioning; V rub Input, V acid and V scratch No input causes the gate voltage of NMOS tube N1 to exceed the threshold voltage, generating leakage current, which corresponds to the animal establishing a connection between acidic stimulation and grasping action in the generalization of classical conditioning, and the part that will also produce salivation when being grasped; V scratch Input, V acid and V rub No input causes the gate voltage of the NMOS tube N1 to exceed the threshold voltage, generating leakage current. This corresponds to the fact that in the generalization of classical conditioning, when the animal is subjected to a scratching action similar to scratching, it will also produce salivation, indicating that generalization is successful.

4. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 1, characterized in that: The acid signal circuit includes: an operational amplifier U1, resistors R1, R8, a switch S1, and a capacitor C1. The same-direction input terminal of U1 is connected to the acid signal, the reverse input terminal of U1 is connected to the output terminal of U1, the left ends of R1 and R2, the right end of R1 is connected to the input terminal of S1, the output terminal of S1 is respectively connected to the left end of C1 and the left end of R8, and the right end of R8 is connected to the right end of R9.

5. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 1, characterized in that: The grasping signal circuit includes: the bending signal circuit includes: resistors R2 to R7, R9, R 10 , R 19 , operational amplifiers U2 to U5 and memory capacitor MC1, the right end of R2 is respectively connected to the right end of R3, the inverting input of U2 and the left end of R4, the left end of R3 is connected to the captured signal, the right end of R4 is respectively connected to the output of U2 and the left end of R5, the right end of R5 is respectively connected to the inverting input of U3 and the left end of R6, the right end of R6 is respectively connected to the output of U3 and the left end of R7, the output of U3 is also connected to the input of S2, the right end of R7 is respectively connected to the left end of MC1 and the same-direction input of U4, the inverting input of U4 is respectively connected to the output of U4 and the left end of R9, the right end of R9 is respectively connected to R 10 The left end of U5 is connected to the reverse input terminal, R 10 The right end is connected to the output end of U5, R 19 The left end of the connection.

6. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 5, characterized in that: The signal circuit of the flexure includes: resistor R 11 To R 18 , R 20 , operational amplifiers U6 to U9, memcapacitor MC2 and switch S2, R 11 The right end of R 12 The right end of U6, the inverting input terminal and R 13 The left end of the connection, R 12 The left end is connected to the torsion signal, R 13 The right end is connected to the output end of U6 and R 14 The left end of the connection, R 14 The right end is connected to the reverse input terminal of U7 and R 15 The left end of the connection, R 15 The right end is connected to the output end of U7, R 16 The left end of the connection, R 16 The right end of is connected to the left end of MC2 and the same-direction input end of U8, and the reverse input end of U8 is connected to the output end of U8 and R 17 The left end of the connection, R 17 The right end of R 18 The left end of the U9 is connected to the reverse input terminal, R 18 The right end is connected to the output end of U9, R 20 The left end of the connection, R 11 The left end of is connected to the output end of S2.

7. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 6, characterized in that: The salivation circuit includes: operational amplifier U 10 , resistor R 21 , R 22 and NMOS tube N1, R 21 The left end of R 19 , R 20 The common right end of U 10 The inverting input terminal is connected to R 21 The right end of 10 The output terminal of N1 is connected to the gate of N1, and the drain of N1 is connected to R 22 The lower end of the connection, R 22 The upper end is connected to a DC voltage.

8. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 7, characterized in that: The model of the operational amplifier is AD8538.

9. The memory capacity classical conditioning reflex generalization differentiation neural network circuit according to claim 7, characterized in that: The model of the NMOS tube is M2SK530.

Images