Disclosure of Invention
Aiming at the defects, the invention provides an energy-saving control method and an energy-saving control system for the super capacitor of the heavy underwater robot, wherein the control strategy unit based on neuron prediction is added into a driving system to predict constant power and peak power required by the motion of the heavy underwater robot in advance, and a rectifier only needs to provide a part of power (lower than rated power) and the rest of power is provided by the optimized super capacitor.
The invention provides the following technical scheme: the energy-saving control method of the super capacitor of the heavy underwater robot is characterized by comprising the following steps of:
s1: for super capacitor, build up with a total resistance Rs And total capacitance CSC The current input into the super capacitor is i (t);
s2: constructing an artificial neural network with an input layer, a hidden layer and an output layer, collecting data indexes of an electric double layer formed by porous electrodes of the supercapacitor, wherein the electric double layer is a first electric layer and a second electric layerA second electric layer having a first complex impedance RC1 The second electric layer has a second complex impedance RC2 Second complex impedance RC1 And a second complex impedance RC2 In series, the electric double layer is preceded by an equivalent capacitance RC0 ；
Calculating t moment by adopting a back propagation method at the hidden layer to generate nonlinear gainAndby means of linear gain-> and />Calculating to obtain an output layer; said nonlinear gain-> and />Are both the operating temperature T (T) of the super capacitor at the time T and the voltage value V of the super capacitor output by the output layerUC (t) relates to;
s3: and feeding back the measurement data obtained by the output layer to the hidden layer, wherein the hidden layer adopts a neuron control strategy method for iteration, and the super capacitor finally obtains an optimal output value through minimum system error correction.
Further, the back propagation method includes the steps of:
1) At time t, a first complex impedance RC1 Capacitance C of (2)1 And resistance R1 Second complex impedance RC2 Capacitance C of (2)2 And resistance R2 Input from an input layer to construct equivalent capacitors RC connected in series in turn0 Capacitance C of (2)0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Total voltage V formedSC Total capacitance C due to operating temperature T as an influencing factorSC Is calculated model C of (2)SC ：
CSC ＝CSC (T,VSC (VC ,V1 ,V2 ))＝Csc (T,X1 ,X2 ,X3 )
wherein ,VC Is equivalent to capacitance RC0 Capacitor C0 The voltage value V1 Voltage formed for capacitance of first complex impedance, V2 Voltage formed for capacitance of second complex impedance, VSC (VC ,V1 ,V2 ) Is equivalent to capacitor C0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Formed total voltage V having nonlinear relationSC ；
wherein ,C0 ×VC 、C1 ×V1 and C2 ×V2 Input values X respectively as input layers1 、X2 and X3 ；
2) Simultaneously constructing equivalent capacitance RC0 Capacitance C of (2)0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Total voltage V formedSC And the total resistance R caused by the operating temperature T as an influencing factorS Is a computational model of (a):
RS ＝RS (T,VSC (VC ,V1 ,V2 ))＝RS (T,X1 ,X2 ,X3 )；
3) Total capacitance C constructed according to step 1) and step 2)SC And total resistance RS And a nonlinear gain generated at time t and />Construction of a nonlinear calculation model at i (t) of the super sensor input currentIs->And output layer results->The output layer results->Based on the nonlinear calculation modelCalculated input value X of input layerk Nonlinear value of the input value at time t +.>Calculating to obtain;for the equivalent capacitance RC0 Capacitance C of (2)0 Nonlinear gain is performed, < >>For the total resistance RS Performing nonlinear gain, wherein ∈>Input value X for the input layer1 、X2 and X3 Average value of>The voltage value V of the super capacitor output by the output layerUC An average value of (t).
Further, the nonlinear calculation model in the step 3)The calculation formula of (2) is as follows:
wherein ,capacitance C being the first complex impedance1 Is a nonlinear form of capacitance value, < >>Capacitance C being the second complex impedance2 Is a non-linear form of capacitance value; />For the resistance R1 Resistance value of nonlinear form of>For the resistance R2 A non-linear form of resistance value of (a); x is Xk (t) is the input value X of the input layerk Input value at time t, < >>Is Xk Nonlinear value of (t), k=1, 2,3.
Further, the output layer result in the step 3)The calculation formula is as follows:
wherein ,is equivalent to capacitance RC0 Capacitance C of (2)0 Is a nonlinear form of capacitance value, < >>Capacitance C being the first complex impedance1 Is a nonlinear form of capacitance value, < >>Capacitance C being the second complex impedance2 Is a non-linear form of capacitance value; />For the total resistance RS Is a non-linear form of resistance value.
Further, the resistor R1 Resistance value of non-linear form of (2)The resistor R2 Resistance value of non-linear form of (2)Capacitance C of the first complex impedance1 Is a nonlinear form of capacitance value->And a capacitance C of the second complex impedance2 Is a nonlinear form of capacitance value->The calculation formulas of (a) are respectively as follows:
wherein , and />Respectively said nonlinear gains +.> and />I is the shorthand for i (t); c (C)UC The supercapacitor capacitance value output for the output layer, < >>Is CUC Is a non-linear capacitance value of (a).
Further, equivalent capacitance RC0 Capacitance C of (2)0 Is a non-linear form of capacitance valueThe calculation formula of (2) isThe total resistance RS Nonlinear form resistance value +.>The calculation formula of (2) is
Further, the total voltage V at time t has a nonlinear relationshipSC The calculation formula of (2) is as follows:
wherein ,to equivalent capacitance RC at time 00 Voltage value of two ends>RC being the first complex impedance at time 01 Voltage value of two ends>To the second complex impedance RC at time 02 Voltage values across the terminals.
The invention also provides an energy-saving control system of the super capacitor of the heavy underwater robot, which comprises a robot power distribution network, a rectifier, a frequency converter, a rotation speed sensor, a propeller, a double DC/DC module, the super capacitor as an energy storage device and a neuron control strategy unit, wherein the robot power distribution network is sequentially connected with the rectifier and the frequency converter, and the double DC/DC module is arranged between the rectifier and the frequency converter.
Further, the other end of the frequency converter is connected with the propeller, and the propeller receives the control of the nerve control strategy unit and transmits the rotating speed data of the propeller to the nerve cell strategy unit to serve as initial data of energy-saving control.
Further, the other end of the bidirectional DC/DC module is connected with the super capacitor serving as the energy storage device, the data output end and the data input end of the super capacitor serving as the energy storage device are connected with the control device in the neuron strategy unit, and an iterative optimization algorithm is performed in the neuron strategy unit to optimize the energy absorbed by the super capacitor serving as the energy storage device and maximize the energy storage of the super capacitor serving as the energy storage device.
The beneficial effects of the invention are as follows:
1. the power and peak power of the driving system of the underwater robot are predicted by adopting a neuron control strategy method to configure the energy storage of the super capacitor in advance, and the starting voltage and the stopping voltage of the super capacitor are obtained by back-pushing, so that the energy storage of the super capacitor is dynamically regulated.
2. The energy consumption of the underwater robot is predicted in advance based on a neuron prediction control strategy, and the super capacitor stores the energy in advance; on the contrary, when the command received by the underwater robot controller shows that the underwater robot is about to move from the water layer to the seabed bottom layer with full power, the underwater robot feeds back larger power in the process, the neuron control strategy unit predicts the energy fed back by the underwater robot in advance and timely adjusts the balance voltage of the super capacitor, so that the super capacitor has enough capacity to absorb the energy fed back in the current process, and the energy which cannot be absorbed by the super capacitor energy storage device is fed back to the power grid through the rectifier in the system, so that the rectifier only needs to process a part of power (lower than rated power), and the power capacity, the switching loss and the harmonic suppression of the rectifier are beneficial.
3. The method provided by the invention adopts a neuron prediction control strategy, can intelligently realize dynamic adjustment of power, and does not need manual intervention; the marine operation system has the advantages of smaller power fluctuation, smaller power supply requirement for the underwater robot mother ship and compatibility of marine operation. The defect that the current high-power propeller system of the underwater robot mostly adopts split working condition control is avoided, and the output power of the propeller in the system can reach a stable state by adjusting the modulation voltage of the electric modulation under the set working condition of the propeller, so that the power consumption is reduced, and the high-efficiency energy-saving control is realized. Meanwhile, peak power of the system can be avoided, and the system is more stable and reliable.
4. The system provided by the invention is applied to a high-power propeller driving system of the heavy-duty robot in water through the neuron control strategy unit, so that the energy-saving control of the heavy-duty robot is realized, and the power capacity of a network is optimized; the voltage balance is adjusted by adjusting the energy storage of the super capacitor serving as the energy storage device, so that peak power under the journey is compensated, a part of rated power can be compensated, and the network power capacity is optimized.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
As shown in fig. 1, the energy-saving control method for the super capacitor of the heavy underwater robot provided by the invention comprises the following steps:
s1: for super capacitor, a capacitor having a total resistance R as shown in FIG. 2 is constructeds And total capacitance CSC The current input into the super capacitor is i (t);
s2: constructing an artificial neural network with an input layer, a hidden layer and an output layer, collecting data indexes of an electric double layer formed by porous electrodes of the super capacitor shown in figure 2, wherein the electric double layer is a first electric layer and a second electric layer, and the first electric layer has a first complex impedance RC1 The second electric layer has a second complex impedance RC2 Second complex impedance RC1 And a second complex impedance RC2 Series, double electricThe layer is provided with an equivalent capacitance RC0 ；
The non-linear gain is generated when the hidden layer adopts a counter propagation method to calculate the moment tAndby means of linear gain-> and />Calculating to obtain an output layer; nonlinear gain-> and />Are the operating temperature T (T) of the super capacitor at the time T and the voltage value V of the super capacitor output by the output layerUC (t) relates to;
s3: and feeding back the measurement data obtained by the output layer to the hidden layer, iterating by adopting a neuron control strategy method by the hidden layer, correcting with the minimum systematic error, and finally obtaining the optimal output value by the supercapacitor.
Example 2
As shown in fig. 1, the energy-saving control method for the super capacitor of the heavy underwater robot provided by the invention comprises the following steps:
s1: for super capacitor, a capacitor having a total resistance R as shown in FIG. 2 is constructeds And total capacitance CSC The current input into the super capacitor is i (t);
s2: constructing an artificial neural network with an input layer, a hidden layer and an output layer, collecting data indexes of an electric double layer formed by porous electrodes of the super capacitor shown in figure 2, wherein the electric double layer is a first electric layer and a second electric layer, and the first electric double layer is a first electric layerThe electric layer has a first complex impedance RC1 The second electric layer has a second complex impedance RC2 Second complex impedance RC1 And a second complex impedance RC2 In series, the electric double layer is previously provided with an equivalent capacitance RC0 ；
The non-linear gain is generated when the hidden layer adopts a counter propagation method to calculate the moment tAndby means of linear gain-> and />Calculating to obtain an output layer; nonlinear gain-> and />Are the operating temperature T (T) of the super capacitor at the time T and the voltage value V of the super capacitor output by the output layerUC (t) relates to;
s3: and feeding back the measurement data obtained by the output layer to the hidden layer, iterating by adopting a neuron control strategy method by the hidden layer, correcting with the minimum systematic error, and finally obtaining the optimal output value by the supercapacitor.
Wherein the back propagation method comprises the steps of:
1) At time t, a first complex impedance RC1 Capacitance C of (2)1 And resistance R1 Second complex impedance RC2 Capacitance C of (2)2 And resistance R2 Input from an input layer to construct equivalent capacitors RC connected in series in turn0 Capacitance C of (2)0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Total voltage V formedSC Work and workTotal capacitance C due to temperature T as influencing factorSC Is calculated model C of (2)SC ：
CSC ＝CSC (T,VSC (VC ,V1 ,V2 ))＝Csc (T,X1 ,X2 ,X3 )
wherein ,VC Is equivalent to capacitance RC0 Capacitance C of (2)0 The voltage value V1 Voltage formed for capacitance of first complex impedance, V2 Voltage formed for capacitance of second complex impedance, VSC (VC ,V1 ,V2 ) Is equivalent to capacitor C0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Formed total voltage V having nonlinear relationSC ；
wherein ,C0 ×VC 、C1 ×V1 and C2 ×V2 Input values X respectively as input layers1 、X2 and X3 ；
2) Simultaneously constructing equivalent capacitance RC0 Capacitance C of (2)0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Total voltage V formedSC And the total resistance R caused by the operating temperature T as an influencing factorS Is a computational model of (a):
RS ＝RS (T,VSC (VC ,V1 ,V2 ))＝RS (T,X1 ,X2 ,X3 )
total voltage V with non-linear relationship at time tSC The calculation formula of (2) is as follows:
wherein ,to equivalent capacitance RC at time 00 Voltage value of two ends>RC as first complex impedance at time 01 Voltage value of two ends>To be at 0 time second complex impedance RC2 The voltage values at both ends;
3) Total capacitance C constructed according to step 1) and step 2)SC And total resistance RS And a nonlinear gain generated at time t and />Construction of a nonlinear calculation model when the input current of the super sensor is i (t)>And output layer results->Output layer results->According to a non-linear calculation model->Calculated input value X of input layerk Nonlinear value of the input value at time t +.>Calculating to obtain; />For equivalent capacitance RC0 Capacitance C of (2)0 Nonlinear gain is performed, < >>For the total resistance RS Performing nonlinear gain, wherein ∈>For input value X of input layer1 、X2 and X3 Average value of>The voltage value V of the super capacitor output by the output layerUC An average value of (t);
wherein the nonlinear calculation modelThe calculation formula of (2) is as follows:
wherein ,capacitor C being a first complex impedance1 Is a nonlinear form of capacitance value, < >>Capacitor C being a second complex impedance2 Is a non-linear form of capacitance value; />Is a resistor R1 Resistance value of nonlinear form of>Is a resistor R2 A non-linear form of resistance value of (a); x is Xk (t) is the input value X of the input layerk Input value at time t, < >>Is Xk Nonlinear value of (t), k=1, 2,3.
Outputting the layer resultThe calculation formula is as follows:
wherein ,is equivalent to capacitance RC0 Capacitance C of (2)0 Is a nonlinear form of capacitance value, < >>Capacitor C being a first complex impedance1 Is a nonlinear form of capacitance value, < >>Capacitor C being a second complex impedance2 Is a non-linear form of capacitance value; />Is the total resistance RS Is a non-linear form of resistance value.
Resistor R1 Resistance value of non-linear form of (2)Resistor R2 Nonlinear form resistance value +.>Capacitance C of first complex impedance1 Is a nonlinear form of capacitance value->And a capacitance C of a second complex impedance2 Is a nonlinear form of capacitance value->Is divided into the calculation formulas of (a)The following are distinguished:
wherein , and />Respectively nonlinear gain-> and />I is the shorthand for i (t); c (C)UC The supercapacitor capacitance value output by the output layer, < +.>Is CUC Is a non-linear capacitance value of (a).
Equivalent capacitance RC0 Capacitance C of (2)0 Is a non-linear form of capacitance valueThe calculation formula of (2) is +.>Total resistance RS Nonlinear form resistance value +.>The calculation formula of (2) is +.>
Example 3
As shown in fig. 1, the energy-saving control method for the super capacitor of the heavy underwater robot provided by the invention comprises the following steps:
s1: for super capacitor, a capacitor having a total resistance R as shown in FIG. 2 is constructeds And total capacitance CSC The current input into the super capacitor is i (t);
s2: constructing an artificial neural network with an input layer, a hidden layer and an output layer, collecting data indexes of an electric double layer formed by porous electrodes of the super capacitor shown in figure 2, wherein the electric double layer is a first electric layer and a second electric layer, and the first electric layer has a first complex impedance RC1 The second electric layer has a second complex impedance RC2 The second electric layer has a second complex impedance RC2 Second complex impedance RC1 And a second complex impedance RC2 In series, the electric double layer is previously provided with an equivalent capacitance RC0 ；
The non-linear gain is generated when the hidden layer adopts a counter propagation method to calculate the moment tAndby means of linear gain-> and />Calculating to obtain an output layer; nonlinear gain-> and />Are the operating temperature T (T) of the super capacitor at the time T and the voltage value V of the super capacitor output by the output layerUC (t) relates to;
s3: and feeding back the measurement data obtained by the output layer to the hidden layer, iterating by adopting a neuron control strategy method by the hidden layer, correcting with the minimum systematic error, and finally obtaining the optimal output value by the supercapacitor.
The neuron control strategy method for iterative optimization has the characteristics of global approximation capability and high success in system identification and complex nonlinear dynamic system control. And the neuron prediction control strategy unit with self-learning and self-adapting capabilities is used for performing high-efficiency energy-saving intelligent control on the power of the underwater robot. The structure is simple, and the device is effectively adapted to the change of the external environment. In order to enhance the high efficiency and robustness of the energy-saving control method, as shown in fig. 3, the neuron control strategy method of the energy-saving control system is established, the rotation speed collected by the propeller is used as the initial data to cooperate with the measurement data obtained by the output layer to carry out assignment iteration, in order to prevent the training of neurons from being interfered when a singular sample exists in the input sample, the weight is adjusted through the Euclidean norm approximation method, and the influence of the singular sample and the rest sample on the weight is in an equilibrium state. After processing by the standardized algorithm, the neuron control strategy method can be described as follows:
in the formula (2), Umax For controlling the maximum limiting value of the quantity, namely the maximum modulation voltage given value of the electric modulation. The selection of the proportionality coefficient K is critical, and the acceleration performance is faster and faster along with the gradual increase of the K value, but the overshoot is increased, so that the transmission system is unstable; if the K value is too small, the system convergence speed will be slow.
The learning algorithm of the neuron control strategy method can be described as:
ωij (k+1)＝ωij (k)+ηij e(k)ui (k)uij (k) (3)
in the equation (3), since the learning rate of the proportional, integral, and differential is the euclidean norm approximation method, the learning efficiency is high. If the overshoot of the system is small and the current process is from overshoot to a steady time process, the sum may be increased; if the overshoot decreases, even below the threshold, then rises to the steady state time course, the value of n may be decreased. Because the neuron control strategy method has a self-learning function, the selection of the initial value of the weighting coefficient does not influence the learning result after the weighting coefficient. The energy-saving control of the designed neuron control strategy unit can be realized by applying the neuron control strategy unit to a high-power propeller driving system.
The power and peak power of the driving system of the underwater robot are predicted by adopting a neuron control strategy method to configure the energy storage of the super capacitor in advance, and the starting voltage and the stopping voltage of the super capacitor are obtained by back-pushing, so that the energy storage of the super capacitor is dynamically regulated. According to the load upper behavior example of the underwater robot, when an instruction received by the underwater robot controller shows that the underwater robot is about to move from the seabed to the water layer in a full load mode, the underwater robot needs to consume larger power in the journey, peak power exists, the energy consumption of the underwater robot is predicted in advance based on a neuron prediction control strategy, and the super capacitor stores the energy in advance; on the contrary, when the command received by the underwater robot controller shows that the underwater robot is about to move from the water layer to the seabed bottom layer with full power, the underwater robot feeds back larger power in the process, the neuron control strategy unit predicts the energy fed back by the underwater robot in advance and timely adjusts the balance voltage of the super capacitor, so that the super capacitor has enough capacity to absorb the energy fed back in the current process, and the energy which cannot be absorbed by the super capacitor energy storage device is fed back to the power grid through the rectifier in the system, so that the rectifier only needs to process a part of power (lower than rated power), and the power capacity, the switching loss and the harmonic suppression of the rectifier are beneficial.
The neuron control strategy unit is applied to a high-power propeller driving system of the heavy-duty robot in water, so that energy-saving control of the heavy-duty robot is realized, and the power capacity of a network is optimized.
Wherein the back propagation method comprises the steps of:
1) At time t, a first complex impedance RC1 Capacitance C of (2)1 And resistance R1 Second complex impedance RC2 Capacitance C of (2)2 And resistance R2 Input from an input layer to construct equivalent capacitors RC connected in series in turn0 Capacitance C of (2)0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Total voltage V formedSC Total capacitance C due to operating temperature T as an influencing factorSC Is calculated model C of (2)SC ：
CSC ＝CSC (T,VSC (VC ,V1 ,V2 ))＝Csc (T,X1 ,X2 ,X3 )
wherein ,VC Is equivalent to capacitance RC0 Capacitance C of (2)0 The voltage value V1 Voltage formed for capacitance of first complex impedance, V2 Voltage formed for capacitance of second complex impedance, VSC (VC ,V1 ,V2 ) Is equivalent to capacitor C0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Formed total voltage V having nonlinear relationSC ；
wherein ,C0 ×VC 、C1 ×V1 and C2 ×V2 Input values X respectively as input layers1 、X2 and X3 ；
2) Simultaneous construction ofEquivalent capacitance RC0 Capacitance C of (2)0 Capacitance C of first complex impedance1 And a capacitance C of a second complex impedance2 Total voltage V formedSC And the total resistance R caused by the operating temperature T as an influencing factorS Is a computational model of (a):
RS ＝RS (T,VSC (VC ,V1 ,V2 ))＝RS (T,X1 ,X2 ,X3 )
total voltage V with non-linear relationship at time tSC The calculation formula of (2) is as follows:
wherein ,to equivalent capacitance RC at time 00 Voltage value of two ends>RC as first complex impedance at time 01 Voltage value of two ends>To be at 0 time second complex impedance RC2 The voltage values at both ends;
3) Total capacitance C constructed according to step 1) and step 2)SC And total resistance RS And a nonlinear gain generated at time t and />Construction of a nonlinear calculation model when the input current of the super sensor is i (t)>And output layer results->Output layer results->According to a non-linear calculation model->Calculated input value X of input layerk Nonlinear value of the input value at time t +.>Calculating to obtain; />For equivalent capacitance RC0 Capacitance C of (2)0 Nonlinear gain is performed, < >>For the total resistance RS Performing nonlinear gain, wherein ∈>For input value X of input layer1 、X2 and X3 Average value of>The voltage value V of the super capacitor output by the output layerUC An average value of (t);
wherein the nonlinear calculation modelThe calculation formula of (2) is as follows:
wherein ,capacitor C being a first complex impedance1 Is a nonlinear form of capacitance value, < >>Capacitor C being a second complex impedance2 Is a non-linear form of capacitance value; />Is a resistor R1 Resistance value of nonlinear form of>Is a resistor R2 A non-linear form of resistance value of (a); x is Xk (t) is the input value X of the input layerk Input value at time t, < >>Is Xk Nonlinear value of (t), k=1, 2,3.
Outputting the layer resultThe calculation formula is as follows:
wherein ,is equivalent to capacitance RC0 Capacitance C of (2)0 Is a nonlinear form of capacitance value, < >>Capacitor C being a first complex impedance1 Is a nonlinear form of capacitance value, < >>Capacitor C being a second complex impedance2 Is a non-linear form of (2)A capacitance value; />Is the total resistance RS Is a non-linear form of resistance value.
Resistor R1 Resistance value of non-linear form of (2)Resistor R2 Nonlinear form resistance value +.>Capacitance C of first complex impedance1 Is a nonlinear form of capacitance value->And a capacitance C of a second complex impedance2 Is a nonlinear form of capacitance value->The calculation formulas of (a) are respectively as follows:
wherein , and />Respectively nonlinear gain-> and />I is the shorthand for i (t); c (C)UC The supercapacitor capacitance value output by the output layer, < +.>Is CUC Is a non-linear capacitance value of (a).
Equivalent capacitance RC0 Capacitance C of (2)0 Is a non-linear form of capacitance valueThe calculation formula of (2) is +.>Total resistance RS Nonlinear form resistance value +.>The calculation formula of (2) is +.>
As shown in fig. 4, the control strategy unit based on neuron prediction provided by the invention is added into the driving system to predict the constant power and peak power required by the motion of the heavy underwater robot in advance, and the rectifier only needs to provide a part of power (lower than rated power) and the rest of power is provided by the super capacitor.
Example 4
As shown in fig. 5, the energy-saving control system of the super capacitor of the heavy underwater robot provided by the invention and adopting the method provided by the embodiments 1-3 comprises a robot power distribution network, a rectifier, a frequency converter, a rotating speed sensor, a propeller, a double DC/DC module, a super capacitor serving as an energy storage device and a neuron control strategy unit, wherein the robot power distribution network is sequentially connected with the rectifier and the frequency converter, and the bidirectional DC/DC module is arranged between the rectifier and the frequency converter; the other end of the frequency converter is connected with a propeller, and the propeller receives the control of the nerve control strategy unit and transmits the rotating speed data of the propeller to the nerve cell strategy unit to serve as initial data of energy-saving control; the other end of the bidirectional DC/DC module is connected with a super capacitor serving as an energy storage device, a data output end and a data input end of the super capacitor serving as the energy storage device are connected with a control device in a neuron strategy unit, an iterative optimization algorithm is performed in the neuron strategy unit, energy absorbed by the super capacitor serving as the energy storage device is optimized, and energy storage of the super capacitor serving as the energy storage device is utilized to the maximum extent.
While the invention has been described with reference to a preferred embodiment, various modifications may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In particular, the technical features mentioned in the respective embodiments may be combined in any manner as long as there is no structural conflict. The present invention is not limited to the specific embodiments disclosed herein, but encompasses all technical solutions falling within the scope of the claims.