Position estimation method, position estimation device, unmanned transport vehicle, and sewing device

Abstract

In one aspect of the position estimation method of the present application, a learning step of acquiring learning data required for estimating the rotational position of the rotor based on the input sensor signals; and a position estimation step of estimating the rotational position of the rotor based on the input sensor signals and the learning data are included. By performing the learning step, data indicating a correspondence relationship between a segment number associated with a division included in each of a plurality of quadrants and a pole pair number indicating a pole pair position is acquired as the learning data. By performing the position estimation step, the initial position of the rotor is determined based on the input sensor signals and the learning data.

Description

Technical Field

[0001] This invention relates to a location estimation method, a location estimation device, an unmanned transport vehicle, and a sewing device. Background Technology

[0002] Previously, electric motors capable of accurately controlling rotor position have been known to incorporate structures with absolute angular position sensors such as optical encoders and rotary transformers. However, absolute angular position sensors are large and costly. Therefore, Patent Document 1 discloses a method for estimating the rotational position of an electric motor rotor without using an absolute angular position sensor.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent No. 6233532 Summary of the Invention

[0006] The technical problem that the invention aims to solve

[0007] In the position estimation method described in Patent Document 1, it is sometimes impossible to estimate the initial position of the rotor rotation within a range where the rotor angle is less than 1 revolution. Therefore, it is difficult to apply to applications where the preparatory action of rotating the rotor to estimate the initial position is not allowed, such as drive motors for robots, unmanned transport vehicles, sewing devices, etc.

[0008] Technical means for solving technical problems

[0009] One aspect of the position estimation method of the present invention is a method for estimating the rotational position of an electric motor comprising a rotor having P (P being an integer greater than or equal to 2) pole pairs, the position estimation method comprising: a learning step for acquiring learning data required to estimate the rotational position; and a position estimation step for estimating the rotational position of the rotor based on the learning data.

[0010] The learning steps include: a first step, wherein a magnet having a single pole pair and sharing a rotation axis with the rotor rotates together with the rotor; a second step, wherein N1 (N1 being an integer of 3 or more) first magnetic sensors, arranged opposite to the magnet and along the rotation direction of the magnet, acquire N1 digital signals whose levels reverse and have a first phase difference with each other when the magnet rotates 180°; a third step, wherein N2 (N2 being an integer of 3 or more) second magnetic sensors, arranged opposite to the rotor and along the rotation direction of the rotor, acquire N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other; and a fourth step, wherein the N... A fifth step involves using one digital signal to divide the learning period into multiple quadrants with N1-bit digital values ​​that are distinct from each other; the fifth step further involves dividing the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, based on the N2 analog signals obtained during the learning period; each of the P pole pair regions is further divided into multiple partitions, and a segment number representing the rotation position is associated with each of the multiple partitions; and a sixth step involves acquiring data representing the correspondence between the segment number and the pole pair number as the learning data, wherein the segment number is associated with the partition contained in each of the multiple quadrants, and the pole pair number represents the pole pair position.

[0011] The position estimation step includes: a 7th step, which acquires the N1 digital signals using the N1 first magnetic sensors; an 8th step, which acquires the N2 analog signals using the N2 second magnetic sensors; a 9th step, which determines the current quadrant from the plurality of quadrants based on the N1 digital signals acquired in the 7th step; a 10th step, which determines the current partition from the plurality of partitions based on the N2 analog signals acquired in the 8th step; and an 11th step, which determines, based on the learning data, the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

[0012] Another aspect of the position estimation method of the present invention is a method for estimating the rotational position of an electric motor including a rotor having P (P being an integer greater than or equal to 2) pole pairs, the position estimation method comprising: a learning step, the learning step acquiring learning data required to estimate the rotational position; and a position estimation step, the position estimation step estimating the rotational position of the rotor based on the learning data.

[0013] The learning steps include: a first step, wherein a magnet having a single pole pair and sharing a rotation axis with the rotor rotates together with the rotor; a second step, wherein N3 (N3 being an integer greater than 2) third magnetic sensors, arranged opposite to the magnet and along the rotation direction of the magnet, acquire N3 analog signals whose electrical signals vary according to the magnetic field strength and have a third phase difference with each other; a third step, wherein N2 (N2 being an integer greater than 3) second magnetic sensors, arranged opposite to the rotor and along the rotation direction of the rotor, acquire N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other; and a fourth step, wherein the fourth step is based on... The fifth step involves using the N3 analog signals obtained during a learning period equivalent to one mechanical angle cycle to calculate the time series data of the mechanical angle during the learning period; the fifth step further involves dividing the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs based on the N2 analog signals obtained during the learning period, further dividing each of the P pole pair regions into multiple partitions, and associating the segment number representing the rotation position with each of the multiple partitions; and the sixth step involves obtaining data representing the correspondence between the time series data of the mechanical angle and the pole pair numbers as the learning data.

[0014] The position estimation step includes: a 7th step, which uses the N3 third magnetic sensors to acquire the N3 analog signals; an 8th step, which calculates the current value of the mechanical angle based on the N3 analog signals acquired in the 7th step; and a 9th step, which determines the pole pair number corresponding to the current value of the mechanical angle as the initial position of the rotor based on the learning data.

[0015] Another aspect of the position estimation method of the present invention is a method for estimating the rotational position of an electric motor including a rotor having P (P being an integer greater than or equal to 2) pole pairs, the position estimation method comprising: a learning step, the learning step acquiring learning data required to estimate the rotational position; and a position estimation step, the position estimation step estimating the rotational position of the rotor based on the learning data.

[0016] The learning steps include: a first step, wherein a magnet having a single pole pair and sharing a rotation axis with the rotor rotates together with the rotor; a second step, wherein N4 (N4 being an integer greater than or equal to 3) fourth magnetic sensors, arranged opposite to the magnet and along the rotation direction of the magnet, acquire N4 analog signals whose electrical signals vary according to the magnetic field strength and have a fourth phase difference with each other; a third step, wherein N2 (N2 being an integer greater than or equal to 3) second magnetic sensors, arranged opposite to the rotor and along the rotation direction of the rotor, acquire N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other; and a fourth step, wherein the fourth step is based on a learning period equivalent to one mechanical angular cycle. The obtained N4 analog signals are used to divide the learning period into multiple quadrants; in step 5, based on the N2 analog signals obtained during the learning period, the learning period is divided into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, each of the P pole pair regions is further divided into multiple partitions, and a segment number representing the rotation position is associated with each of the multiple partitions; and in step 6, data representing the correspondence between the segment number and the pole pair number is obtained as the learning data, wherein the segment number is associated with the partition contained in each of the multiple quadrants, and the pole pair number represents the pole pair position.

[0017] The location estimation step has the following characteristics:

[0018] Step 7, which uses the N4 fourth magnetic sensors to acquire the N4 analog signals; Step 8, which uses the N2 second magnetic sensors to acquire the N2 analog signals; Step 9, which determines the current quadrant from the plurality of quadrants based on the N4 analog signals acquired in Step 7; Step 10, which determines the current partition from the plurality of partitions based on the N2 analog signals acquired in Step 8; and Step 11, which determines, based on the learning data, the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

[0019] One aspect of the position estimation device of the present invention is a device for estimating the rotational position of an electric motor including a rotor having P (P being an integer of 2 or more) magnetic pole pairs. The position estimation device includes: a magnet having one magnetic pole pair and sharing a rotation axis with the rotor; N1 (N1 being an integer of 3 or more) first magnetic sensors, the N1 first magnetic sensors being opposite to the magnet and arranged along the rotational direction of the magnet; N2 (N2 being an integer of 3 or more) second magnetic sensors, the N2 second magnetic sensors being opposite to the rotor and arranged along the rotational direction of the rotor; and a signal processing device that processes the output signals of the first magnetic sensors and the second magnetic sensors.

[0020] The signal processing apparatus includes: a processing unit that performs learning processing for acquiring learning data required to predict the rotational position, and a position prediction processing for predicting the rotational position of the rotor based on the learning data; and a storage unit that stores the learning data.

[0021] As part of the learning process, the processing unit performs: a first process, which causes the magnet to rotate together with the rotor; a second process, which acquires N1 digital signals with reversed levels and a first phase difference from each other every 180° rotation of the magnet using the N1 first magnetic sensors; a third process, which acquires N2 analog signals with varying electrical signals according to the magnetic field strength and a second phase difference from each other using the N2 second magnetic sensors; a fourth process, which divides the learning period into multiple quadrants with N1 distinct digital values ​​based on the N1 digital signals acquired during a learning period equivalent to one mechanical angle cycle; and a fifth process. The fifth process, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions; and the sixth process stores data representing the correspondence between the segment number and the pole pair number as the learning data in the storage unit, wherein the segment number is associated with the partition contained in each of the multiple quadrants, and the pole pair number represents the pole pair position.

[0022] As part of the position estimation process, the processing unit performs: a seventh process, which acquires the N1 digital signals using the N1 first magnetic sensors; an eighth process, which acquires the N2 analog signals using the N2 second magnetic sensors; a ninth process, which determines the current quadrant from the plurality of quadrants based on the N1 digital signals acquired in the seventh process; a tenth process, which determines the current partition from the plurality of partitions based on the N2 analog signals acquired in the eighth process; and an eleventh process, which determines, based on the learning data, the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

[0023] Another aspect of the position estimation device of the present invention is a device for estimating the rotational position of an electric motor including a rotor having P (P being an integer of 2 or more) magnetic pole pairs. The position estimation device includes: a magnet having one magnetic pole pair and sharing a rotation axis with the rotor; N3 (N3 being an integer of 2 or more) third magnetic sensors, the N3 third magnetic sensors being opposite to the magnet and configured along the rotational direction of the magnet; N2 (N2 being an integer of 3 or more) second magnetic sensors, the N2 second magnetic sensors being opposite to the rotor and configured along the rotational direction of the rotor; and a signal processing device that processes the output signals of the second magnetic sensors and the third magnetic sensors.

[0024] The signal processing apparatus includes: a processing unit that performs learning processing for acquiring learning data required to predict the rotational position, and position prediction processing for predicting the rotational position of the rotor based on the learning data; and a storage unit that stores the learning data.

[0025] As part of the learning process, the processing unit performs: a first process, which causes the magnet to rotate together with the rotor; a second process, which acquires N3 analog signals whose electrical signals vary according to the magnetic field strength and have a third phase difference with each other through the N3 third magnetic sensors; a third process, which acquires N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other through the N2 second magnetic sensors; and a fourth process, which calculates the value of the learning process based on the N3 analog signals acquired during a learning period equivalent to one mechanical angular cycle. The fifth process, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions; and the sixth process, which stores data representing the correspondence between the time series data of the mechanical angle and the pole pair numbers as the learning data in the storage unit.

[0026] As part of the position estimation process, the processing unit performs: a seventh process, which acquires the N3 analog signals using the N3 third magnetic sensors; an eighth process, which calculates the current value of the mechanical angle based on the N3 analog signals acquired in the seventh process; and a ninth process, which determines the pole pair number corresponding to the current value of the mechanical angle as the initial position of the rotor based on the learning data stored in the storage unit.

[0027] Another aspect of the position estimation device of the present invention is a device for estimating the rotational position of an electric motor including a rotor having P (P being an integer of 2 or more) magnetic pole pairs. The position estimation device includes: a magnet having one magnetic pole pair and sharing a rotation axis with the rotor; N4 (N4 being an integer of 3 or more) fourth magnetic sensors, the N4 fourth magnetic sensors being opposite to the magnet and configured along the rotational direction of the magnet; N2 (N2 being an integer of 3 or more) second magnetic sensors, the N2 second magnetic sensors being opposite to the rotor and configured along the rotational direction of the rotor; and a signal processing device that processes the output signals of the second magnetic sensors and the fourth magnetic sensors.

[0028] The signal processing apparatus includes: a processing unit that performs learning processing to acquire learning data required to infer the rotational position, and a position inference processing for inferring the rotational position of the rotor based on the learning data; and a storage unit that stores the learning data.

[0029] As part of the learning process, the processing unit performs: a first process, which causes the magnet to rotate together with the rotor; a second process, which acquires N4 analog signals whose electrical signals vary according to the magnetic field strength and have a fourth phase difference with each other through the N4 fourth magnetic sensors; a third process, which acquires N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other through the N2 second magnetic sensors; a fourth process, which divides the learning period into multiple quadrants based on the N4 analog signals acquired during a learning period equivalent to one mechanical angular cycle; and a fifth process, which... The learning period is divided into P pole pair regions by the N2 analog signals obtained during the learning period and each of the P pole pair regions is further divided into multiple partitions, and a segment number representing the rotation position is associated with each of the multiple partitions; and a sixth process is performed to store data representing the correspondence between the segment number and the pole pair number as the learning data in the storage unit, wherein the segment number is associated with the partition contained in each of the multiple quadrants, and the pole pair number represents the pole pair position.

[0030] As part of the position estimation process, the processing unit performs: a seventh process, which acquires the N4 analog signals using the N4 fourth magnetic sensors; an eighth process, which acquires the N2 analog signals using the N2 second magnetic sensors; a ninth process, which determines the current quadrant from the plurality of quadrants based on the N4 analog signals acquired in the seventh process; a tenth process, which determines the current partition from the plurality of partitions based on the N2 analog signals acquired in the eighth process; and an eleventh process, which determines, based on the learning data, the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

[0031] One aspect of the unmanned transport vehicle of the present invention includes: an electric motor having a rotor having P (P being an integer greater than or equal to 2) pole pairs; and a position estimation device for estimating the rotational position of the electric motor according to any one of the above three aspects.

[0032] One aspect of the sewing apparatus of the present invention includes: an electric motor having a rotor having P (P being an integer greater than or equal to 2) pole pairs; and a position estimation device for estimating the rotational position of the electric motor according to any one of the above three aspects.

[0033] Invention Effects

[0034] According to the above aspects of the present invention, a position estimation method, a position estimation device, an unmanned transport vehicle, and a sewing device are provided, which can eliminate the need for a preparatory rotational action for estimating the rotational position. Attached Figure Description

[0035] Figure 1 This is a block diagram schematically illustrating the structure of the position estimation device according to Embodiment 1 of the present invention.

[0036] Figure 2 This is a flowchart showing the learning process performed by the processing unit in Embodiment 1.

[0037] Figure 3 This is an explanatory diagram regarding the learning data obtained through the learning process in Implementation 1.

[0038] Figure 4 It is an enlarged view of the incremental signals Hu, Hv, and Hw contained within a pole pair region.

[0039] Figure 5 This is a flowchart showing the position estimation process performed by the processing unit in Embodiment 1.

[0040] Figure 6 This is a diagram showing a variation of the position estimation device in Embodiment 1.

[0041] Figure 7 This is a block diagram schematically illustrating the structure of the position estimation device according to Embodiment 2 of the present invention.

[0042] Figure 8 This is a flowchart showing the learning process performed by the processing unit in Embodiment 2.

[0043] Figure 9 This is an explanatory diagram regarding the learning data obtained through the learning process in Implementation Method 2.

[0044] Figure 10 This is a flowchart showing the position estimation process performed by the processing unit in Embodiment 2.

[0045] Figure 11 This is a block diagram schematically illustrating the structure of the position estimation device according to Embodiment 3 of the present invention.

[0046] Figure 12This is a flowchart showing the learning process performed by the processing unit in Embodiment 3.

[0047] Figure 13 This is an explanatory diagram regarding the learning data obtained through the learning process in Implementation Method 3.

[0048] Figure 14 This is a flowchart showing the position estimation process performed by the processing unit in Embodiment 3.

[0049] Figure 15 This is a diagram showing a variation of embodiment 3, example 1.

[0050] Figure 16 This is a diagram showing a variation of embodiment 3, example 2.

[0051] Figure 17 This is a diagram showing the appearance of an unmanned transport vehicle as an application example of the present invention.

[0052] Figure 18 This is a diagram showing the appearance of a sewing device as an application example of the present invention. Detailed Implementation

[0053] Hereinafter, with reference to the accompanying drawings, one embodiment of the present invention will be described in detail.

[0054] [Implementation Method 1]

[0055] Figure 1 This is a block diagram schematically illustrating the structure of the position estimation device 100 according to Embodiment 1 of the present invention. Figure 1 As shown, the position estimation device 100 is a device for estimating the rotational position (rotation angle) of a motor 200, which includes a rotor 210 having P (P being an integer greater than or equal to 2) pole pairs. In this embodiment, as an example, the rotor 210 has four pole pairs. Furthermore, a pole pair refers to a pair of N and S poles. That is, in this embodiment, the rotor 210 has four pairs of N and S poles, totaling eight poles (rotor magnets).

[0056] Motor 200 is, for example, an internal rotor type three-phase brushless DC motor. Although in Figure 1 (Illustrations omitted) The motor 200, in addition to the rotor 210, also includes a stator and a motor housing. The motor housing internally houses the rotor 210 and the stator. The rotor 210 is rotatably supported about a rotation axis by bearing components inside the motor housing. The stator has three-phase excitation coils including U-phase coils, V-phase coils, and W-phase coils, and is fixed inside the motor housing in a state opposite to the outer peripheral surface of the rotor 210.

[0057] The position estimation device 100 includes a sensor magnet 10, three first magnetic sensors 21, 22, and 23, three second magnetic sensors 31, 32, and 33, and a signal processing device 40. Although in Figure 1 The diagram is omitted, but a circuit board is installed in the motor 200, and the first magnetic sensors 21, 22, 23, the second magnetic sensors 31, 32, 33, and the signal processing device 40 are disposed on the circuit board.

[0058] The sensor magnet 10 is a circular plate-shaped magnet with one pair of magnetic poles and sharing a rotation axis with the rotor 210. When the rotor 210 rotates, the sensor magnet 10 rotates synchronously with the rotor 210. The sensor magnet 10 is positioned in a location that does not interfere with the circuit board. The sensor magnet 10 can be disposed inside the motor housing or outside the motor housing.

[0059] The first magnetic sensors 21, 22, and 23 are magnetic sensors arranged at predetermined intervals on a circuit board opposite to the sensor magnet 10 and along the rotation direction of the sensor magnet 10. This embodiment illustrates a position estimation device 100 including three first magnetic sensors 21, 22, and 23, but the number of first magnetic sensors can be N1 (N1 is an integer greater than or equal to 3). For example, the first magnetic sensors 21, 22, and 23 are Hall effect ICs that incorporate Hall elements and latching circuits. The first magnetic sensors 21, 22, and 23 each output a digital signal, the level of which is reversed every 180° rotation of the sensor magnet 10.

[0060] In this embodiment, the first magnetic sensors 21, 22, and 23 are arranged at 120° intervals along the rotation direction of the sensor magnet 10. Therefore, the digital signals output from the first magnetic sensors 21, 22, and 23 have a 120° phase difference (first phase difference) with each other in electrical angle. Hereinafter, the digital signals output from the first magnetic sensors 21, 22, and 23 will be referred to as absolute digital signals. The first magnetic sensor 21 outputs an absolute digital signal HA1 to the signal processing device 40. The first magnetic sensor 22 outputs an absolute digital signal HA2 to the signal processing device 40. The first magnetic sensor 23 outputs an absolute digital signal HA3 to the signal processing device 40.

[0061] The second magnetic sensors 31, 32, and 33 are magnetic sensors arranged at predetermined intervals on the circuit board opposite to the rotor 210 and along the rotation direction of the rotor 210. This embodiment illustrates a position estimation device 100 including three second magnetic sensors 31, 32, and 33, but the number of second magnetic sensors can be N² (N² is an integer greater than or equal to 3). For example, the second magnetic sensors 31, 32, and 33 are Hall elements or linear Hall ICs, respectively. The second magnetic sensors 31, 32, and 33 each output analog signals whose electrical signals vary according to the magnetic field strength. One electrical angular period of each analog signal corresponds to 1 / P of one mechanical angular period. In this embodiment, since the number of pole pairs P of the rotor 210 is "4", one electrical angular period of each analog signal corresponds to 1 / 4 of one mechanical angular period, i.e., a mechanical angle of 90°.

[0062] In this embodiment, the second magnetic sensors 31, 32, and 33 are arranged at 30° intervals along the rotation direction of the rotor 210. Therefore, the analog signals output from the second magnetic sensors 31, 32, and 33 have a 120° phase difference (second phase difference) with each other in electrical angle. Hereinafter, the analog signals output from the second magnetic sensors 31, 32, and 33 will be referred to as incremental signals. The second magnetic sensor 31 outputs the incremental signal Hu to the signal processing device 40. The second magnetic sensor 32 outputs the incremental signal Hv to the signal processing device 40. The second magnetic sensor 33 outputs the incremental signal Hw to the signal processing device 40.

[0063] The signal processing device 40 is used to process the output signals of the first magnetic sensors 21, 22, and 23 and the second magnetic sensors 31, 32, and 33. The signal processing device 40 infers the rotational position of the motor 200, i.e., the rotational position of the rotor 210, based on the absolute digital signals HA1, HA2, and HA3 and the incremental signals Hu, Hv, and Hw. The signal processing device 40 includes a processing unit 41 and a storage unit 42.

[0064] The processing unit 41 is a microprocessor, such as an MCU (Microcontroller Unit). Absolute digital signals HA1, HA2, HA3 and incremental signals Hu, Hv, Hw are input to the processing unit 41. The processing unit 41 is connected to the storage unit 42 via a data bus in a manner that enables data communication.

[0065] Furthermore, inside the processing unit 41, the incremental signals Hu, Hv, and Hw are converted into digital signals via an A / D converter. For ease of explanation, the digital signals output from the A / D converter are also referred to as the incremental signals Hu, Hv, and Hw. In the following description, the absolute digital signals HA1, HA2, and HA3, as well as the incremental signals Hu, Hv, and Hw input to the processing unit 41, are sometimes collectively referred to as "input sensor signals".

[0066] The processing unit 41 performs at least the following two processes according to the program stored in the storage unit 42: The processing unit 41 performs learning processing to acquire learning data required to predict the rotational position of the rotor 210 based on input sensor signals. The processing unit 41 performs position prediction processing to predict the rotational position of the rotor 210 based on the input sensor signals and the learning data.

[0067] The storage unit 42 includes: a non-volatile memory storing programs, various settings, and learning data required for the processing unit 41 to perform various processes; and a volatile memory used as a temporary storage destination for data when the processing unit 16 performs various processes. The non-volatile memory is, for example, EEPROM (Electrically Erasable Programmable Read-Only Memory) or flash memory. The volatile memory is, for example, RAM (Random Access Memory).

[0068] Next, the learning process performed by the processing unit 41 will be explained. The learning process corresponds to the learning step in the position estimation method of the first aspect. Figure 2 This is a flowchart illustrating the learning process performed by the processing unit 41 in Embodiment 1. When the power to the signal processing device 40 is first turned on, the processing unit 41 performs... Figure 2 The signal processing device 40 performs at least the learning process based on the learning step and the position inference step.

[0069] like Figure 2 As shown, if the learning process begins, the processing unit 41 first executes a first process (step S1) that causes the sensor magnet 10 to rotate together with the rotor 210. This first process corresponds to the first step of the learning step in the position estimation method of the first aspect.

[0070] Next, the processing unit 41 performs a second process (step S2) to acquire three absolute digital signals HA1, HA2, and HA3 through the three first magnetic sensors 21, 22, and 23. This second process corresponds to the second step of the learning step in the position estimation method of the first aspect.

[0071] like Figure 3As shown, the absolute digital signals HA1, HA2, and HA3 are digital signals whose levels are reversed every 180° of rotation of the sensor magnet 10, and which have a 120° phase difference with each other in electrical angle. Figure 3 In this context, the period from time t1 to time t9 corresponds to one mechanical angular period. Figure 3 In this context, the periods from time t1 to time t2, from time t2 to time t4, from time t4 to time t5, from time t5 to time t6, from time t6 to time t8, and from time t8 to time t9 are each equivalent to 1 / 6 of a mechanical angle period, i.e., a mechanical angle of 60°.

[0072] Next, the processing unit 41 performs a third process (step S3) to acquire three incremental signals Hu, Hv, and Hw through the three second magnetic sensors 31, 32, and 33. This third process corresponds to the third step of the learning step in the position estimation method of the first aspect.

[0073] like Figure 3 As shown, one electrical angular period of each of the incremental signals Hu, Hv, and Hw is equivalent to 1 / 4 of one mechanical angular period, i.e., the mechanical angle is 90°. Figure 3 In this context, the periods from time t1 to time t3, from time t3 to time t5, from time t5 to time t7, and from time t7 to time t9 each correspond to 90° in mechanical angles. Furthermore, the incremental signals Hu, Hv, and Hw have a 120° phase difference from each other in electrical angles.

[0074] Next, the processing unit 41 performs a fourth process (step S4) that divides the learning period into multiple quadrants with N1-bit digital values ​​that are different from each other, based on the three absolute digital signals HA1, HA2, and HA3 obtained during the learning period equivalent to one mechanical angle cycle. This fourth process corresponds to the fourth step of the learning step in the position estimation method of the first aspect.

[0075] "N1" represents the number of the first magnetic sensors. Therefore, in this embodiment, since there are three first magnetic sensors, in step S4, the processing unit 41 divides the learning period into multiple quadrants with distinct 3-bit digital values. In this embodiment, among the 3-bit digital values, the value of the topmost bit is the value of the absolute digital signal HA1, the value of the middle bit is the value of the absolute digital signal HA2, and the value of the bottommost bit is the value of the absolute digital signal HA3.

[0076] like Figure 3 As shown, the processing unit 41 divides the learning period (one mechanical angular cycle) into 6 quadrants based on the absolute digital signals HA1, HA2, and HA3.

[0077] The processing unit 41 divides the learning period from time t1 to time t2 into the first quadrant, which has a 3-bit digital value "101".

[0078] The processing unit 41 divides the learning period from time t2 to time t4 into the second quadrant, which has a 3-bit digital value "100".

[0079] The processing unit 41 divides the learning period from time t4 to time t5 into the third quadrant, which has a 3-bit digital value "110".

[0080] The processing unit 41 divides the learning period from time t5 to time t6 into the fourth quadrant, which has a 3-bit digital value "010".

[0081] The processing unit 41 divides the learning period from time t6 to time t8 into the fifth quadrant, which has a 3-bit digital value "011".

[0082] The processing unit 41 divides the learning period from time t8 to time t9 into the sixth quadrant, which has a 3-bit digital value "001".

[0083] Next, the processing unit 41 performs a fifth process, which, based on the three incremental signals Hu, Hv, and Hw obtained during the learning period, divides the learning period into four pole pair regions associated with pole pair numbers representing the pole pair positions of each of the four pole pairs. Each of the four pole pair regions is further divided into multiple partitions, and a segment number representing the rotational position of the rotor 210 is associated with each of the multiple partitions (step S5). This fifth process corresponds to the fifth step of the learning steps in the position estimation method of the first aspect.

[0084] In this embodiment, to determine the rotational position of the rotor 210, pole pair numbers representing pole pair positions are assigned to the four magnetic pole pairs of the rotor 210. For example, as Figure 1 As shown, for the four magnetic pole pairs of rotor 210, the pole pair numbers are assigned in a clockwise direction in the order of "0", "1", "2", "3".

[0085] like Figure 3 As shown, in step S5, the processing unit 41 divides the learning period into four pole pair regions based on the three incremental signals Hu, Hv, and Hw obtained during the learning period. Figure 3 In this context, "No.C" indicates the pole pair number.

[0086] The processing unit 41 divides the period from time t1 to time t3 during the learning period into pole pair regions associated with the pole pair number "0".

[0087] The processing unit 41 divides the period from time t3 to time t5 during the learning period into pole pair regions associated with the pole pair number "1".

[0088] The processing unit 41 divides the period from time t5 to time t7 during the learning period into pole pair regions associated with the pole pair number "2".

[0089] The processing unit 41 divides the period from time t7 to time t9 during the learning period into pole pair regions associated with the pole pair number "3".

[0090] like Figure 3 As shown, in step S5, the processing unit 41, based on the three incremental signals Hu, Hv, and Hw obtained during the learning period, further divides each of the four pole pair regions into 12 partitions, and associates the segment number representing the rotational position of the rotor 210 with each of the 12 partitions. Figure 3 In this context, "No.A" represents the partition number assigned to the partition, and "No.B" represents the segment number.

[0091] like Figure 3 As shown, each of the 12 partitions within the four pole pairs is assigned a partition number from "0" to "11". On the other hand, consecutive numbers throughout the learning period are associated with each partition as segment numbers. Specifically, as... Figure 3 As shown, in the pole pair region associated with pole pair number “0”, the segment numbers “0” to “11” are associated with the partition numbers “0” to “11”.

[0092] In the pole pair region associated with pole pair number “1”, section numbers “12” to “23” are associated with section numbers “0” to “11”.

[0093] In the pole pair region associated with pole pair number "2", segment numbers "24" to "35" are associated with partition numbers "0" to "11". In the pole pair region associated with pole pair number "3", segment numbers "36" to "47" are associated with partition numbers "0" to "11".

[0094] Figure 4 This is a magnified view of the incremental signals Hu, Hv, and Hw within a single pole pair region. See below for reference. Figure 4 Specifically, the fifth process executed by processing unit 41 includes the process of dividing each of the four pole-pair regions in the process into 12 partitions. Figure 4 In this context, the baseline value for amplitude is "0". Figure 4 In the example, a positive amplitude value represents the numerical value of the magnetic field strength at the N pole. A negative amplitude value represents the numerical value of the magnetic field strength at the S pole.

[0095] In the fifth step of the learning process, the processing unit 41 performs a process to extract zero-crossing points, which are the points where the three incremental signals Hu, Hv, and Hw included in each of the four pole pair regions intersect with the reference value "0". For example... Figure 4 As shown, the processing unit 41 extracts points P1, P3, P5, P7, P9, P11, and P13 as zero-crossing points.

[0096] In the fifth step of the learning process, processing unit 41 performs a process to extract intersection points, which are the points where the three incremental signals Hu, Hv, and Hw included in each of the four pole pair regions intersect each other. For example... Figure 4 As shown, the processing unit 41 extracts points P2, P4, P6, P8, P10, and P12 as intersection points.

[0097] In the fifth process of learning, the processing unit 41 performs a process that determines the interval between adjacent zero-crossing points and intersection points as partitions.

[0098] like Figure 4 As shown, the processing unit 41 determines the interval between the zero point P1 and the intersection point P2 as the partition with the partition number "0".

[0099] The processing unit 41 determines the interval between the intersection point P2 and the zero-crossing point P3 as the partition with the partition number "1".

[0100] The processing unit 41 determines the interval between the zero-crossing point P3 and the intersection point P4 as the partition to be assigned partition number "2".

[0101] The processing unit 41 determines the interval between the intersection point P4 and the zero-crossing point P5 as the partition with the partition number "3".

[0102] The processing unit 41 determines the interval between the zero point P5 and the intersection point P6 as the partition with the partition number "4".

[0103] The processing unit 41 determines the interval between the intersection point P6 and the zero-crossing point P7 as the partition with the partition number "5".

[0104] The processing unit 41 determines the interval between the zero point P7 and the intersection point P8 as the partition with the partition number "6".

[0105] The processing unit 41 determines the interval between the intersection point P8 and the zero-crossing point P9 as the partition to be assigned partition number "7".

[0106] The processing unit 41 determines the interval between the zero point P9 and the intersection point P10 as the partition with the partition number "8".

[0107] The processing unit 41 determines the interval between the intersection point P10 and the zero-crossing point P11 as the partition with the partition number "9".

[0108] The processing unit 41 determines the interval between the zero point P11 and the intersection point P12 as the partition with the partition number "10".

[0109] The processing unit 41 determines the interval between the intersection point P12 and the zero-crossing point P13 as the partition to be assigned partition number "11".

[0110] Through the fifth process of the above learning process, such as Figure 3 As shown, during the learning process, the system is divided into four pole pair regions associated with pole pair numbers. Each of the four pole pair regions is further divided into 12 partitions, with the segment number associated with each partition. Additionally, in the following description, for example, the partition assigned partition number "0" will be referred to as "partition 0," and the partition assigned partition number "11" will be referred to as "partition 11."

[0111] Next, the processing unit 41 performs a sixth process: acquiring data representing the correspondence between segment numbers and pole pair numbers as learning data and storing the acquired learning data in the storage unit 42. The segment number is associated with a partition contained in each of the six quadrants, and the pole pair number represents the pole pair position (step S6). This sixth process corresponds to the sixth step of the learning steps in the position estimation method of the first aspect.

[0112] like Figure 3 As shown, Quadrant 1 contains eight partitions, from partition 0 to partition 7. The learning data includes data representing the correspondence between the segment numbers "0" to "7" associated with partitions 0 to 7 contained in Quadrant 1, and the pole pair number "0".

[0113] like Figure 3 As shown, Quadrant 2 contains four partitions, numbered 8 through 11, and four partitions, numbered 0 through 3. The training data includes data representing the correspondence between segment numbers "8" to "11" associated with partitions 8 through 11 in Quadrant 2 and the pole pair number "0". Additionally, it includes data representing the correspondence between segment numbers "12" to "15" associated with partitions 0 through 3 in Quadrant 2 and the pole pair number "1".

[0114] like Figure 3 As shown, quadrant 3 contains eight partitions, from partition 4 to partition 11. The learning data includes data representing the correspondence between the segment numbers "16" to "23" associated with partitions 4 to 11 contained in quadrant 3, and the pole pair number "1".

[0115] like Figure 3As shown, quadrant 4 contains eight partitions, from partition 0 to partition 7. The learning data includes data representing the correspondence between the segment numbers "24" to "31" associated with partitions 0 to 7 in quadrant 4, and the pole pair number "2".

[0116] like Figure 3 As shown, quadrant 5 contains four partitions, numbered 8 through 11, and four partitions, numbered 0 through 3. The training data includes data representing the correspondence between segment numbers "32" through "35" associated with partitions 8 through 11 in quadrant 5 and the pole pair number "2". Additionally, it includes data representing the correspondence between segment numbers "36" through "39" associated with partitions 0 through 3 in quadrant 5 and the pole pair number "3".

[0117] like Figure 3 As shown, quadrant 6 contains eight partitions, from partition 4 to partition 11. The learning data includes data representing the correspondence between the segment numbers "40" to "47" associated with partitions 4 to 11 contained in quadrant 6, and the pole pair number "3".

[0118] Next, the position estimation process performed by the processing unit 41 will be explained. The position estimation process corresponds to the position estimation step in the position estimation method of the first aspect. Figure 5 This is a flowchart illustrating the position estimation process performed by the processing unit 41 in Embodiment 1. After the above-described learning process is performed, when the power to the signal processing device 40 is re-energized, the processing unit 41 performs... Figure 5 The location is inferred as shown.

[0119] like Figure 5 As shown, if the position estimation process begins, the processing unit 41 first performs the seventh process (step S7) to acquire three absolute digital signals HA1, HA2, and HA3 through the three first magnetic sensors 21, 22, and 23. This seventh process corresponds to the seventh step of the position estimation steps in the position estimation method of the first aspect.

[0120] Next, the processing unit 41 performs an eighth process (step S8) to acquire three incremental signals Hu, Hv, and Hw through the three second magnetic sensors 31, 32, and 33. This eighth process corresponds to the eighth step of the position estimation step in the position estimation method of the first aspect.

[0121] Next, the processing unit 41 performs a ninth process (step S9) to determine the current quadrant from the six quadrants based on the three absolute digital signals HA1, HA2, and HA3 obtained in the seventh process (step S7) described above. This ninth process corresponds to the ninth step of the position estimation step in the position estimation method of the first aspect.

[0122] In step S9, the processing unit 41 determines the current quadrant from the six quadrants based on the 3-bit digital values ​​represented by the absolute digital signals HA1, HA2, and HA3. For example, if the 3-bit digital value is "100", the processing unit 41 determines the second quadrant as the current quadrant.

[0123] Next, the processing unit 41 executes a 10th process (step S10) to determine the current partition from the 12 partitions based on the three incremental signals Hu, Hv, and Hw obtained in the 8th process (step S8) described above. This 10th process corresponds to the 10th step of the position estimation step in the position estimation method of the first aspect.

[0124] In step S10, the processing unit 41 determines the current partition from the 12 partitions, for example, based on the magnitude relationship of the detection values ​​of the incremental signals Hu, Hv, and Hw, and the positive or negative sign of each detection value. Figure 4 As shown, for example, in partition 2, the detected value of the incremental signal Hu is the largest and has a positive sign. Furthermore, in partition 2, the detected value of the incremental signal Hw is the second largest and has a negative sign. Additionally, in partition 2, the detected value of the incremental signal Hv is the smallest and has a negative sign. When the magnitude relationship of the detected values ​​of the incremental signals Hu, Hv, and Hw, and the sign of each detected value, satisfy the conditions for partition 2 described above, the processing unit 41 determines partition 2 as the current partition.

[0125] Next, the processing unit 41 performs the eleventh process, which, based on the learning data stored in the storage unit 42, determines the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant as the initial position of the rotor 210 (step S11). This eleventh process corresponds to the eleventh step of the position estimation step in the position estimation method of the first aspect.

[0126] For example, as mentioned above, assuming quadrant 2 is determined as the current quadrant, partition 2 is determined as the current partition. Figure 3As shown, the learning data includes data representing the correspondence between segment numbers "12" to "15" associated with partitions 0 to 3 contained in the second quadrant and the pole pair number "1". Therefore, when the second quadrant is determined as the current quadrant and partition 2 is determined as the current partition, the processing unit 41 determines the pole pair number "1" corresponding to segment number "14" as the initial position of the rotor 210, which is associated with partition 2 contained in the second quadrant.

[0127] As described above, the position estimation device 100 of Embodiment 1 includes a processing unit 41, which performs learning processing to acquire learning data required to estimate the rotational position of the rotor 210 based on input sensor signals, and position estimation processing to estimate the rotational position of the rotor 210 based on the input sensor signals and the learning data. The processing unit 41 performs the learning processing at least when the power to the signal processing device 40 is first turned on, thereby acquiring data representing the correspondence between segment numbers and pole pair numbers as learning data. The segment number is associated with a partition contained in each of the six quadrants, and the pole pair number represents the pole pair position. The processing unit 41 determines the initial position of the rotor 210 by performing the position estimation processing when the power to the signal processing device 40 is re-turned on.

[0128] Therefore, the position estimation device 100 of Embodiment 1 can estimate the initial position of the rotor 210 without causing the rotor 210 to rotate. Thus, when the power is turned on, the motor 200, including the position estimation device 100, does not need to adjust the origin of the rotor 210's rotational position. The motor 200 does not require a preparatory rotational action for origin adjustment, and therefore can be appropriately used as a drive motor for robots, automated guided vehicles, etc., where a preparatory rotational action is not permitted. Since the motor 200 does not require a preparatory rotational action for origin adjustment, the drive time and power consumption required for the preparatory rotational action can be reduced.

[0129] (A variation of Implementation Method 1)

[0130] The present invention is not limited to the above-described embodiment 1. The various structures described in this specification can be appropriately combined within a range that does not contradict each other.

[0131] In Embodiment 1, examples are shown where the first magnetic sensors 21, 22, and 23 are Hall effect ICs that incorporate Hall elements and latching circuits. For example, such as Figure 6As shown, a comparator circuit 44 can be provided in the signal processing device 40. This comparator circuit 44 replaces the first magnetic sensors 21, 22, and 23 with Hall elements, respectively, and converts the analog signals output from the three Hall elements into absolute digital signals HA1, HA2, and HA3. The comparator circuit 44 can be provided outside the processing unit 41 or inside the processing unit 41.

[0132] Furthermore, in the above embodiment 1, the case of setting three first magnetic sensors that output absolute digital signals is illustrated. However, the number of first magnetic sensors is not limited to three. The number of first magnetic sensors can be N1 (N1 is an integer greater than or equal to 3).

[0133] Furthermore, in the above embodiment 1, the case of setting a second magnetic sensor with 3 output incremental signals is shown. However, the number of the second magnetic sensors is not limited to 3. The number of the second magnetic sensors can be N2 (N2 is an integer greater than or equal to 3).

[0134] Furthermore, in the above embodiment 1, an electric motor including a rotor with 4 pole pairs was illustrated, but the number of pole pairs of the rotor is not limited to 4, and the number of pole pairs of the rotor can be P (P is an integer greater than or equal to 2).

[0135] [Implementation Method 2]

[0136] Next, Embodiment 2 of the present invention will be described.

[0137] Figure 7 This is a block diagram schematically illustrating the structure of the position estimation device 110 according to Embodiment 2 of the present invention. Figure 7 As shown, the position estimation device 110 is a device for estimating the rotational position (rotation angle) of a motor 200, which includes a rotor 210 having P pole pairs (P being an integer greater than or equal to 2). In this embodiment, as an example, the rotor 210 has four pole pairs. Since the structure of the motor 200 is the same as in Embodiment 1, the description of the motor 200 is omitted in Embodiment 2.

[0138] The position estimation device 110 includes a sensor magnet 10, two third magnetic sensors 51 and 52, three second magnetic sensors 31, 32, and 33, and a signal processing device 40. Although in Figure 7 The diagram is omitted, but a circuit board is installed in the motor 200, and the third magnetic sensors 51 and 52, the second magnetic sensors 31, 32 and 33, and the signal processing device 40 are disposed on the circuit board.

[0139] The sensor magnet 10 is the same as in Embodiment 1. That is, the sensor magnet 10 is a circular plate-shaped magnet having a single pair of magnetic poles and sharing a rotation axis with the rotor 210. When the rotor 210 rotates, the sensor magnet 10 rotates synchronously with the rotor 210.

[0140] The third magnetic sensors 52 and 53 are magnetic sensors arranged on the circuit board opposite to the sensor magnet 10 and at predetermined intervals along the rotation direction of the sensor magnet 10. This embodiment illustrates a position estimation device 110 including two third magnetic sensors 51 and 52, but the number of third magnetic sensors can be N³ (N³ is an integer greater than or equal to 2). For example, the third magnetic sensors 51 and 52 are Hall elements or linear Hall ICs, respectively. The third magnetic sensors 51 and 52 each output analog signals whose electrical signals vary according to the magnetic field strength. One electrical angular cycle of the analog signals output from the third magnetic sensors 51 and 52 corresponds to one mechanical angular cycle.

[0141] In this embodiment, the third magnetic sensors 51 and 52 are arranged at 90° intervals along the rotation direction of the sensor magnet 10. Therefore, the analog signals output from the third magnetic sensors 51 and 52 have a 90° phase difference (third phase difference) with each other in electrical angle. Hereinafter, the analog signals output from the third magnetic sensors 51 and 52 will be referred to as absolute analog signals. The third magnetic sensor 51 outputs the absolute analog signal HB1 to the signal processing device 40. The third magnetic sensor 52 outputs the absolute analog signal HB2 to the signal processing device 40.

[0142] The second magnetic sensors 31, 32, and 33 are magnetic sensors arranged on the circuit board opposite to the rotor 210 and at predetermined intervals along the rotation direction of the rotor 210. Since the second magnetic sensors 31, 32, and 33 are the same as in Embodiment 1, the description of the second magnetic sensors 31, 32, and 33 in Embodiment 2 is omitted.

[0143] The signal processing apparatus 40 in Embodiment 2 infers the rotational position of the motor 200, i.e., the rotational position of the rotor 210, based on the absolute analog signals HB1 and HA2 output from the third magnetic sensors 51 and 52, and the incremental signals Hu, Hv, and Hw output from the second magnetic sensors 31, 32, and 33. The signal processing apparatus 40 includes a processing unit 41 and a storage unit 42. The storage unit 42 is the same as in Embodiment 1, therefore, the description of the storage unit 42 in Embodiment 2 is omitted.

[0144] Absolute analog signals HB1, HB2 and incremental signals Hu, Hv, Hw are input to processing unit 41. Inside processing unit 41, absolute analog signals HB1, HB2 and incremental signals Hu, Hv, Hw are converted into digital signals by an A / D converter. For ease of explanation, the digital signals output from the A / D converter are also referred to as absolute analog signals HB1, HB2 and incremental signals Hu, Hv, Hw. Furthermore, in the following description, the absolute analog signals HB1, HB2 and incremental signals Hu, Hv, Hw input to processing unit 41 are sometimes collectively referred to as "input sensor signals".

[0145] The processing unit 41 performs at least the following two processes according to the program stored in the storage unit 42: The processing unit 41 performs a learning process to acquire learning data required to predict the rotational position of the rotor 210 based on the input sensor signals. The processing unit 41 performs a position prediction process to predict the rotational position of the rotor 210 based on the input sensor signals and the learning data. In Embodiment 2, the content of the learning process and the position prediction process performed by the processing unit 41 differs from that in Embodiment 1.

[0146] Next, the learning process performed by the processing unit 41 of Embodiment 2 will be described. The learning process corresponds to the learning step in the position estimation method of the second aspect. Figure 8 This is a flowchart illustrating the learning process performed by the processing unit 41 in Embodiment 2. The processing unit 41 performs this process at least when the signal processing device 40 is first powered on. Figure 8 The learning process is shown below.

[0147] like Figure 8 As shown, if the learning process begins, the processing unit 41 first executes the first process (step S21) that causes the sensor magnet 10 to rotate together with the rotor 210. This first process corresponds to the first step of the learning step in the position estimation method of the second aspect.

[0148] Next, the processing unit 41 performs a second process (step S22) to acquire two absolute analog signals HB1 and HB2 via the two third magnetic sensors 51 and 52. This second process corresponds to the second step of the learning step in the position estimation method of the second aspect.

[0149] like Figure 9 As shown, one electrical angular period of each of the absolute analog signals HB1 and HB2 is equivalent to one mechanical angular period. Figure 9 In this context, the period from time t1 to time t9 corresponds to one mechanical angular cycle. Furthermore, the absolute analog signals HB1 and HB2 have a 90° phase difference with each other in electrical angle.

[0150] Next, the processing unit 41 performs a third process (step S23) to acquire three incremental signals Hu, Hv, and Hw through the three second magnetic sensors 31, 32, and 33. This third process corresponds to the third step of the learning step in the position estimation method of the second aspect.

[0151] like Figure 9 As shown, similar to Embodiment 1, one electrical angular period of each of the incremental signals Hu, Hv, and Hw corresponds to 1 / 4 of one mechanical angular period, i.e., a mechanical angle of 90°. Figure 3 In this context, the periods from time t1 to time t3, from time t3 to time t5, from time t5 to time t7, and from time t7 to time t9 each correspond to 90° in mechanical angles. Furthermore, the incremental signals Hu, Hv, and Hw have a 120° phase difference from each other in electrical angles.

[0152] Next, the processing unit 41 performs a fourth process (step S24) to calculate the time series data of the mechanical angle θ during the learning period based on the two absolute analog signals HB1 and HB2 obtained during the learning period, which is equivalent to one mechanical angle cycle. This fourth process corresponds to the fourth step of the learning step in the position estimation method of the second aspect.

[0153] For example, in step S24, the processing unit 41 samples the absolute analog signals HB1 and HB2 obtained during the learning period at a predetermined sampling frequency, and substitutes the sampled values ​​of the absolute analog signals HB1 and HB2 into the following calculation formula (1) to calculate the time series data of the machine angle θ. Hereinafter, the time series data of the machine angle θ will be referred to as the machine angle time series data.

[0154] Mechanical angle θ = tan -1 (HB1 / HB2)…(1)

[0155] Next, the processing unit 41 performs a fifth process, which, based on the three incremental signals Hu, Hv, and Hw obtained during the learning period, divides the learning period into four pole pair regions associated with pole pair numbers representing the pole pair positions of each of the four pole pairs. Each of the four pole pair regions is further divided into multiple partitions, and a segment number representing the rotational position of the rotor 210 is associated with each of the multiple partitions (step S25). This fifth process corresponds to the fifth step of the learning steps in the position estimation method of the second aspect.

[0156] By performing the process in step S25, such as Figure 9As shown, during the learning process, the system is divided into four pole pair regions associated with pole pair numbers. Each of the four pole pair regions is further divided into 12 partitions, and the segment number is associated with each partition. The processing in step S25 is the same as step S5 in the learning process of Embodiment 1, therefore the description of step S25 in Embodiment 2 is omitted.

[0157] Next, the processing unit 41 performs a sixth process, which acquires data representing the correspondence between the machine angle time series data and the pole pair number as learning data, and stores the acquired learning data in the storage unit 42 (step S26). This sixth process corresponds to the sixth step of the learning step in the position estimation method of the second aspect.

[0158] For example, such as Figure 9 As shown, in the training data, pole pair number "0" corresponds to the mechanical angle θ from 0° (360°) to 89° in the mechanical angle time series data. In the training data, pole pair number "1" corresponds to the mechanical angle θ from 90° to 179° in the mechanical angle time series data. In the training data, pole pair number "2" corresponds to the mechanical angle θ from 180° to 269° in the mechanical angle time series data. In the training data, pole pair number "3" corresponds to the mechanical angle θ from 270° to 359° in the mechanical angle time series data.

[0159] Next, the position estimation process performed by the processing unit 41 of Embodiment 2 will be described. The position estimation process of Embodiment 2 corresponds to the position estimation step in the position estimation method of the second aspect. Figure 10 This is a flowchart illustrating the position estimation process performed by the processing unit 41 in Embodiment 2. After execution... Figure 8 After the learning process shown, when the power to the signal processing device 40 is reconnected, the processing unit 41 executes... Figure 10 The location is inferred as shown.

[0160] like Figure 10 As shown, if the position estimation process begins, the processing unit 41 first performs the seventh process (step S27) to acquire two absolute analog signals HB1 and HA2 using the two third magnetic sensors 51 and 52. This seventh process corresponds to the seventh step of the position estimation steps in the position estimation method of the second aspect.

[0161] Next, the processing unit 41 executes an eighth process (step S28) to calculate the current value of the machine angle θ based on the two absolute analog signals HB1 and HA2 obtained in the seventh process (step S27). This eighth process corresponds to the eighth step of the position estimation step in the position estimation method of the second aspect. In step S28, the processing unit 41 calculates the current value of the machine angle θ by substituting the sampled values ​​of the absolute analog signals HB1 and HB2 into the above-described formula (1).

[0162] Next, the processing unit 41 performs a ninth process (step S29) to determine the pole pair number corresponding to the current value of the mechanical angle θ as the initial position of the rotor 210 based on the learning data stored in the storage unit 42. This ninth process corresponds to the ninth step of the position estimation step in the position estimation method of the second aspect.

[0163] For example, suppose the current value of the machine angle θ is calculated as a value within the range of 90° to 179°. As described above, in the learning data, the pole pair number "1" corresponds to the machine angle θ from 90° to 179° in the machine angle time series data. Therefore, when the current value of the machine angle θ is calculated as a value within the range of 90° to 179°, the processing unit 41 determines the pole pair number "1" corresponding to the current value of the machine angle θ as the initial position of the rotor 210.

[0164] As described above, the position estimation device 110 of Embodiment 2 includes a processing unit 41, which performs learning processing to acquire learning data required to estimate the rotational position of the rotor 210 based on input sensor signals, and position estimation processing to estimate the rotational position of the rotor 210 based on the input sensor signals and the learning data. The processing unit 41 performs the learning processing at least when the power to the signal processing device 40 is first turned on, thereby acquiring data representing the correspondence between mechanical angle time series data and pole pair numbers as learning data. The processing unit 41 determines the initial position of the rotor 210 by performing position estimation processing when the power to the signal processing device 40 is re-turned on.

[0165] Therefore, similar to Embodiment 1, the position estimation device 110 of Embodiment 2 can estimate the initial position of the rotor 210 without causing the rotor 210 to rotate. Thus, when the power is turned on, the motor 200 including the position estimation device 110 does not need to adjust the origin of the rotor 210's rotational position. The motor 200 does not require a preparatory rotational action for origin adjustment, and therefore can also be appropriately used as a drive motor for robots, automated guided vehicles, etc., where a preparatory rotational action is not permitted. Since the motor 200 does not require a preparatory rotational action for origin adjustment, the drive time and power consumption required for the preparatory rotational action can be reduced.

[0166] (A variation of Implementation Method 2)

[0167] The present invention is not limited to the above-described embodiment 2. The various structures described in this specification can be appropriately combined within a range that does not contradict each other.

[0168] In the above embodiment 2, an example was shown where two third magnetic sensors with absolute analog output signals were provided. However, the number of third magnetic sensors is not limited to two; the number of third magnetic sensors can be N3 (N3 is an integer greater than or equal to 2). That is, the number of third magnetic sensors can be three or more.

[0169] Furthermore, in the above embodiment 2, the case of setting a second magnetic sensor with 3 output incremental signals is shown. However, the number of the second magnetic sensors is not limited to 3. The number of the second magnetic sensors can be N2 (N2 is an integer greater than or equal to 3).

[0170] Furthermore, in the above embodiment 2, an electric motor including a rotor with 4 pole pairs was illustrated, but the number of pole pairs of the rotor is not limited to 4, and the number of pole pairs of the rotor can be P (P is an integer greater than or equal to 2).

[0171] [Implementation Method 3]

[0172] Next, Embodiment 3 of the present invention will be described.

[0173] Figure 11 This is a block diagram schematically illustrating the structure of the position estimation device 120 according to Embodiment 3 of the present invention. Figure 11 As shown, the position estimation device 120 is a device for estimating the rotational position (rotation angle) of a motor 200, which includes a rotor 210 having P pole pairs (P being an integer greater than or equal to 2). In this embodiment, as an example, the rotor 210 has four pole pairs. Since the structure of the motor 200 is the same as in Embodiment 1, the description of the motor 200 is omitted in Embodiment 3.

[0174] The position estimation device 120 includes a sensor magnet 10, three fourth magnetic sensors 61, 62, and 63, three second magnetic sensors 31, 32, and 33, and a signal processing device 40. Although in Figure 11 The diagram is omitted, but a circuit board is installed in the motor 200, and the fourth magnetic sensors 61, 62, and 63, the second magnetic sensors 31, 32, and 33, and the signal processing device 40 are disposed on the circuit board.

[0175] The sensor magnet 10 is the same as in Embodiment 1. That is, the sensor magnet 10 is a circular plate-shaped magnet having a single pair of magnetic poles and sharing a rotation axis with the rotor 210. When the rotor 210 rotates, the sensor magnet 10 rotates synchronously with the rotor 210.

[0176] The fourth magnetic sensors 61, 62, and 63 are magnetic sensors arranged on the circuit board opposite to the sensor magnet 10 and at predetermined intervals along the rotation direction of the sensor magnet 10. This embodiment illustrates a position estimation device 120 including three fourth magnetic sensors 61, 62, and 63, but the number of fourth magnetic sensors can be N⁴ (N⁴ is an integer greater than or equal to 3). For example, the fourth magnetic sensors 61, 62, and 63 are Hall elements or linear Hall ICs, respectively. The fourth magnetic sensors 61, 62, and 63 each output analog signals whose electrical signals vary according to the magnetic field strength. One electrical angular cycle of the analog signal output from each of the fourth magnetic sensors 61, 62, and 63 corresponds to one mechanical angular cycle.

[0177] In this embodiment, the fourth magnetic sensors 61, 62, and 63 are arranged at 120° intervals along the rotation direction of the sensor magnet 10. Therefore, the analog signals output from the fourth magnetic sensors 61, 62, and 63 have a 120° phase difference (fourth phase difference) with each other in electrical angle. Hereinafter, the analog signals output from the fourth magnetic sensors 61, 62, and 63 will be referred to as absolute analog signals. The fourth magnetic sensor 61 outputs an absolute analog signal HC1 to the signal processing device 40. The fourth magnetic sensor 62 outputs an absolute analog signal HC2 to the signal processing device 40. The fourth magnetic sensor 63 outputs an absolute analog signal HC3 to the signal processing device 40.

[0178] The second magnetic sensors 31, 32, and 33 are magnetic sensors arranged on the circuit board opposite to the rotor 210 and at predetermined intervals along the rotation direction of the rotor 210. Since the second magnetic sensors 31, 32, and 33 are the same as in Embodiment 1, the description of the second magnetic sensors 31, 32, and 33 in Embodiment 3 is omitted.

[0179] The signal processing apparatus 40 in Embodiment 3 infers the rotational position of the motor 200, i.e., the rotational position of the rotor 210, based on the absolute analog signals HC1, HC2, HC3 output from the fourth magnetic sensors 61, 62, 63 and the incremental signals Hu, Hv, Hw output from the second magnetic sensors 31, 32, 33. The signal processing apparatus 40 includes a processing unit 41 and a storage unit 42. The storage unit 42 is the same as in Embodiment 1, therefore, the description of the storage unit 42 in Embodiment 3 is omitted.

[0180] Absolute analog signals HC1, HC2, HC3 and incremental signals Hu, Hv, Hw are input to processing unit 41. Inside processing unit 41, absolute analog signals HC1, HC2, HC3 and incremental signals Hu, Hv, Hw are converted into digital signals by an A / D converter. For ease of explanation, the digital signals output from the A / D converter are also referred to as absolute analog signals HC1, HC2, HC3 and incremental signals Hu, Hv, Hw. Furthermore, in the following description, the absolute analog signals HC1, HC2, HC3 and incremental signals Hu, Hv, Hw input to processing unit 41 are sometimes collectively referred to as "input sensor signals".

[0181] The processing unit 41 performs at least the following two processes according to the program stored in the storage unit 42: The processing unit 41 performs a learning process to acquire learning data required to predict the rotational position of the rotor 210 based on input sensor signals. The processing unit 41 performs a position prediction process to predict the rotational position of the rotor 210 based on the input sensor signals and the learning data. In Embodiment 3, the content of the learning process and the position prediction process performed by the processing unit 41 differs from that in Embodiments 1 and 2.

[0182] Next, the learning process performed by the processing unit 41 of Embodiment 3 will be described. The learning process corresponds to the learning step in the position estimation method of the third aspect. Figure 12 This is a flowchart illustrating the learning process performed by the processing unit 41 in Embodiment 3. The processing unit 41 performs this process at least when the signal processing device 40 is first powered on. Figure 12 The learning process is shown below.

[0183] like Figure 12 As shown, if the learning process begins, the processing unit 41 first executes a first process (step S41) that causes the sensor magnet 10 to rotate together with the rotor 210. This first process corresponds to the first step of the learning process in the position estimation method of the third aspect.

[0184] Next, the processing unit 41 performs a second process (step S42) to acquire three absolute analog signals HC1, HC2, and HC3 through the three fourth magnetic sensors 61, 62, and 63. This second process corresponds to the second step of the learning step in the position estimation method of the third aspect.

[0185] like Figure 13 As shown, one electrical angular period of each of the absolute analog signals HC1, HC2, and HC3 is equivalent to one mechanical angular period. Figure 13 In this context, the period from time t1 to time t9 corresponds to one mechanical angular cycle. Furthermore, the absolute analog signals HC1, HC2, and HC3 have a 120° phase difference with each other in electrical angle.

[0186] Next, the processing unit 41 performs a third process (step S43) to acquire three incremental signals Hu, Hv, and Hw through the three second magnetic sensors 31, 32, and 33. This third process corresponds to the third step of the learning step in the position estimation method of the third aspect.

[0187] like Figure 13 As shown, similar to Embodiment 1, one electrical angular period of each of the incremental signals Hu, Hv, and Hw corresponds to 1 / 4 of one mechanical angular period, i.e., a mechanical angle of 90°. Figure 13 In this context, the periods from time t1 to time t3, from time t3 to time t5, from time t5 to time t7, and from time t7 to time t9 each correspond to 90° in mechanical angles. Furthermore, the incremental signals Hu, Hv, and Hw have a 120° phase difference from each other in electrical angles.

[0188] Next, the processing unit 41 performs a fourth process (step S44) that divides the learning period into multiple quadrants based on the three absolute analog signals HC1, HC2, and HC3 obtained during the learning period, which is equivalent to one mechanical angle cycle. This fourth process corresponds to the fourth step of the learning step in the position estimation method of the third aspect.

[0189] In step S44, the processing unit 41 performs a process to extract the zero-crossing point, which is the point where the three absolute analog signals HC1, HC2, and HC3 intersect with the reference value "0". For example... Figure 13 As shown, the processing unit 41 extracts points P20, P21, P22, P23, P24, P25, and P26 as the zero-crossing points of the absolute analog signals HC1, HC2, and HC3. Then, the processing unit 41 divides the interval between two adjacent zero-crossing points into quadrants.

[0190] like Figure 13 As shown, the processing unit 41 divides the interval between the zero-crossing point P20 and the zero-crossing point P21 into the first quadrant.

[0191] The processing unit 41 divides the interval between the zero-crossing point P21 and the zero-crossing point P22 into the second quadrant.

[0192] The processing unit 41 divides the interval between the zero-crossing point P22 and the zero-crossing point P23 into the third quadrant.

[0193] The processing unit 41 divides the interval between the zero-crossing point P23 and the zero-crossing point P24 into the fourth quadrant.

[0194] The processing unit 41 divides the interval between the zero-crossing point P24 and the zero-crossing point P25 into the fifth quadrant.

[0195] The processing unit 41 divides the interval between the zero-crossing point P25 and the zero-crossing point P26 into the 6th quadrant.

[0196] Therefore, in Embodiment 3, the processing unit 41 divides the learning period into six quadrants based on the absolute analog signals HC1, HC2, and HC3.

[0197] Next, the processing unit 41 performs a fifth process, which, based on the three incremental signals Hu, Hv, and Hw obtained during the learning period, divides the learning period into four pole pair regions associated with pole pair numbers representing the pole pair positions of each of the four pole pairs. Each of the four pole pair regions is further divided into multiple partitions, and a segment number representing the rotational position of the rotor 210 is associated with each of the multiple partitions (step S45). This fifth process corresponds to the fifth step of the learning steps in the position estimation method of the third aspect.

[0198] By performing the process in step S45, such as Figure 13 As shown, during the learning process, the system is divided into four pole pair regions associated with pole pair numbers. Each of the four pole pair regions is further divided into 12 partitions, and the segment number is associated with each partition. The processing in step S45 is the same as step S5 in the learning process of embodiment 1, therefore the description of step S45 in embodiment 3 will be omitted.

[0199] Next, the processing unit 41 performs a sixth process: acquiring data representing the correspondence between segment numbers and pole pair numbers as learning data and storing the acquired learning data in the storage unit 42. The segment number is associated with a partition contained in each of the six quadrants, and the pole pair number represents the pole pair position (step S46). This sixth process corresponds to the sixth step of the learning steps in the position estimation method of the third aspect. Figure 13 As shown, since the learning data obtained in Embodiment 3 is the same as the learning data obtained in Embodiment 1, the description of the learning data in Embodiment 3 is omitted.

[0200] Next, the position estimation process performed by the processing unit 41 of Embodiment 3 will be described. The position estimation process of Embodiment 3 corresponds to the position estimation step in the position estimation method of the third aspect. Figure 14 This is a flowchart illustrating the position estimation process performed by the processing unit 41 in Embodiment 3. After execution... Figure 12 After the learning process shown, when the power to the signal processing device 40 is reconnected, the processing unit 41 executes... Figure 14 The location is inferred as shown.

[0201] like Figure 14As shown, if the position estimation process begins, the processing unit 41 first performs the seventh process (step S7) to acquire three absolute analog signals HC1, HC2, and HC3 using the three fourth magnetic sensors 61, 62, and 63. This seventh process corresponds to the seventh step of the position estimation steps in the position estimation method of the third aspect.

[0202] Next, the processing unit 41 performs the eighth process (step S48) by acquiring three incremental signals Hu, Hv, and Hw through the three second magnetic sensors 31, 32, and 33.

[0203] This eighth process corresponds to the eighth step of the location estimation steps in the location estimation method of the third aspect.

[0204] Next, the processing unit 41 performs a ninth process (step S9) to determine the current quadrant from the six quadrants based on the three absolute analog signals HC1, HC2, and HC3 obtained in the seventh process (step S47) described above. This ninth process corresponds to the ninth step of the position estimation step in the position estimation method of the third aspect.

[0205] In step S49, the processing unit 41 determines the current quadrant from the six quadrants, for example, based on the magnitude relationship of the detection values ​​of the absolute analog signals HC1, HC2, and HC3, and the positive or negative sign of each detection value. For example, suppose the processing unit 41 determines the second quadrant as the current quadrant from the six quadrants.

[0206] Next, the processing unit 41 executes a 10th process (step S50) to determine the current partition from the 12 partitions based on the three incremental signals Hu, Hv, and Hw obtained in the 8th process (step S48) described above. This 10th process corresponds to the 10th step of the position estimation step in the position estimation method of the third aspect. The method for determining the current partition is the same as in Embodiment 1, therefore the description of the method for determining the current partition in Embodiment 3 is omitted. For example, assume that the processing unit 41 determines partition number 2 as the current partition.

[0207] Next, the processing unit 41 performs an eleventh process, which, based on the learning data stored in the storage unit 42, determines the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant as the initial position of the rotor 210 (step S51). This eleventh process corresponds to the eleventh step of the position estimation step in the position estimation method of the third aspect.

[0208] For example, as mentioned above, assuming quadrant 2 is determined as the current quadrant, partition 2 is determined as the current partition. Figure 13As shown, the learning data includes data representing the correspondence between segment numbers "12" to "15" associated with partitions 0 to 3 contained in the second quadrant and the pole pair number "1". Therefore, when the second quadrant is determined as the current quadrant and partition 2 is determined as the current partition, the processing unit 41 determines the pole pair number "1" corresponding to the segment number "14" associated with partition 2 contained in the second quadrant as the initial position of the rotor 210.

[0209] As described above, the position estimation device 120 of Embodiment 3 includes a processing unit 41, which performs learning processing to acquire learning data required to estimate the rotational position of the rotor 210 based on input sensor signals, and position estimation processing to estimate the rotational position of the rotor 210 based on the input sensor signals and the learning data. The processing unit 41 performs the learning processing at least when the power to the signal processing device 40 is first turned on, thereby acquiring data representing the correspondence between segment numbers and pole pair numbers as learning data. The segment number is associated with a partition contained in each of the six quadrants, and the pole pair number represents the pole pair position. The processing unit 41 determines the initial position of the rotor 210 by performing the position estimation processing when the power to the signal processing device 40 is re-turned on.

[0210] Therefore, similar to Embodiment 1, the position estimation device 120 in Embodiment 3 can estimate the initial position of the rotor 210 without causing the rotor 210 to rotate. Thus, when the power is turned on, the motor 200, including the position estimation device 120, does not need to adjust the origin of the rotor 210's rotational position. Since the motor 200 does not require a preparatory rotational action for origin adjustment, it can also be appropriately used as a drive motor for robots, automated guided vehicles, etc., where a preparatory rotational action is not permitted. Because the motor 200 does not require a preparatory rotational action for origin adjustment, the drive time and power consumption required for the preparatory rotational action can be reduced.

[0211] (A variation of implementation method 3)

[0212] The present invention is not limited to the above-described embodiment 3. The various structures described in this specification can be appropriately combined within a range that does not contradict each other.

[0213] In the above-described embodiment 3, an example is illustrated where the processing unit 41 divides the learning period into six quadrants by extracting the zero-crossing points of the three absolute analog signals HC1, HC2, and HC3 obtained during a learning period equivalent to one mechanical angle cycle. For example, as... Figure 15 As shown, the processing unit 41 can divide the learning period into six quadrants by extracting the intersection of the three absolute analog signals HC1, HC2, and HC3 obtained during the learning period, which is equivalent to one mechanical angle cycle.

[0214] Specifically, in Figure 15 In the modified example shown, the processing unit 41 performs the process of extracting the intersection points, i.e., the points where the three absolute analog signals HC1, HC2, and HC3 intersect. For example... Figure 15 As shown, the processing unit 41 extracts points P27, P28, P29, P30, P31, and P32 as the intersection points of the absolute analog signals HC1, HC2, and HC3. Then, the processing unit 41 divides the interval between any two adjacent intersection points into quadrants.

[0215] like Figure 15 As shown, processing unit 41 divides the interval between intersection point P32 and intersection point P27 into the first quadrant. Processing unit 41 divides the interval between intersection point P27 and intersection point P28 into the second quadrant. Processing unit 41 divides the interval between intersection point P28 and intersection point P29 into the third quadrant. Processing unit 41 divides the interval between intersection point P29 and intersection point P30 into the fourth quadrant. Processing unit 41 divides the interval between intersection point P30 and intersection point P31 into the fifth quadrant. Processing unit 41 divides the interval between intersection point P31 and intersection point P32 into the sixth quadrant.

[0216] For example, such as Figure 16 As shown, the processing unit 41 can divide the learning period into 12 quadrants by extracting the zero-crossing points and intersection points of the three absolute analog signals HC1, HC2, and HC3 obtained during a learning period equivalent to one mechanical angle cycle. Specifically, in Figure 16 In the modified example shown, the processing unit 41 divides the interval between adjacent zero-crossing points and intersection points into quadrants.

[0217] like Figure 16 As shown, processing unit 41 divides the interval between the zero-crossing point P20 and the intersection point P27 into the first quadrant. Processing unit 41 divides the interval between the intersection point P27 and the zero-crossing point P21 into the second quadrant. Processing unit 41 divides the interval between the zero-crossing point P21 and the intersection point P28 into the third quadrant. Processing unit 41 divides the interval between the intersection point P28 and the zero-crossing point P22 into the fourth quadrant. Processing unit 41 divides the interval between the zero-crossing point P22 and the intersection point P29 into the fifth quadrant. Processing unit 41 divides the interval between the intersection point P29 and the zero-crossing point P23 into the sixth quadrant.

[0218] Processing unit 41 divides the interval between the zero-crossing point P23 and the intersection point P30 into the 7th quadrant. Processing unit 41 divides the interval between the intersection point P30 and the zero-crossing point P24 into the 8th quadrant. Processing unit 41 divides the interval between the zero-crossing point P24 and the intersection point P31 into the 9th quadrant. Processing unit 41 divides the interval between the intersection point P31 and the zero-crossing point P25 into the 10th quadrant. Processing unit 41 divides the interval between the zero-crossing point P25 and the intersection point P32 into the 11th quadrant. Processing unit 41 divides the interval between the intersection point P32 and the zero-crossing point P26 into the 12th quadrant.

[0219] Therefore, the learning period can be divided into 12 quadrants by extracting the zero-crossing points and intersection points of the three absolute analog signals HC1, HC2, and HC3 obtained during the learning period, which is equivalent to one mechanical angle cycle.

[0220] Furthermore, in the above embodiment 3, a case was illustrated where three fourth magnetic sensors with absolute analog output signals were provided. However, the number of fourth magnetic sensors is not limited to three; the number of fourth magnetic sensors can be N4 (N4 is an integer greater than or equal to 3). In other words, the number of fourth magnetic sensors can be four or more.

[0221] Furthermore, in the above embodiment 3, the case of setting a second magnetic sensor with 3 output incremental signals is shown. However, the number of the second magnetic sensors is not limited to 3. The number of the second magnetic sensors can be N2 (N2 is an integer greater than or equal to 3).

[0222] Furthermore, in the above embodiment 3, an electric motor including a rotor with 4 pole pairs was illustrated, but the number of pole pairs of the rotor is not limited to 4, and the number of pole pairs of the rotor can be P (P is an integer greater than or equal to 2).

[0223] [Application Example]

[0224] Figure 17 A diagram showing the appearance of an unmanned transport vehicle 300, which is an application example of the present invention. Figure 18 A diagram showing the appearance of a sewing device 400 as an application example of the present invention is provided.

[0225] The unmanned transport vehicle 300 and the sewing device 400 include an electric motor having a rotor with P (P being an integer greater than or equal to 2) pole pairs; and a position estimation device that estimates the rotational position of the electric motor. The electric motor 200 described in the above embodiments can be used as the electric motor. Alternatively, any one of the position estimation devices described in Embodiment 1 (position estimation device 100), Embodiment 2 (position estimation device 110), and Embodiment 3 (position estimation device 120) can be used as the position estimation device. The electric motors in the unmanned transport vehicle 300 and the sewing device 400 do not require a preparatory rotational movement for origin adjustment, thus preventing unwanted movements of the unmanned transport vehicle 300 and the sewing device 400. Furthermore, the application examples of this invention are not limited to the unmanned transport vehicle 300 and the sewing device 400; this invention can be widely applied to devices such as robots where a preparatory rotational movement of the electric motor is not permissible.

[0226] Label Explanation

[0227] 100, 110, 120… Position estimation device, 10… Sensor magnet (magnet), 21, 22, 23… First magnetic sensor, 31, 32, 33… Second magnetic sensor, 40… Signal processing device, 41… Processing unit, 42… Storage unit, 51, 52… Third magnetic sensor, 61, 62, 63… Fourth magnetic sensor, 200… Electric motor, 210… Rotor, 300… Unmanned transport vehicle, 400… Sewing device.

Claims

1. A position estimation method for estimating the rotational position of an electric motor comprising a rotor having P pole pairs, where P is an integer greater than or equal to 2, the position estimation method being characterized by having: The learning steps involve acquiring learning data needed to infer the rotational position; and The position estimation step, based on the learned data, estimates the rotational position of the rotor. The learning steps have the following characteristics: The first step involves rotating a magnet having a pole pair and sharing a rotation axis with the rotor together with the rotor. In the second step, N1 first magnetic sensors, which are arranged opposite to the magnet and along the rotation direction of the magnet, are used to acquire N1 digital signals whose levels are reversed and have a first phase difference with each other when the magnet rotates 180°. N1 is an integer greater than or equal to 3. The third step utilizes N2 second magnetic sensors arranged opposite to the rotor and along the rotation direction of the rotor to acquire N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other, where N2 is an integer greater than or equal to 3. The fourth step, based on the N1 digital signals obtained during a learning period equivalent to one mechanical angular cycle, divides the learning period into multiple quadrants with N1-bit digital values ​​that are different from each other. The fifth step, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions. as well as Step 6 involves acquiring data representing the correspondence between the segment number and the pole pair number, as the learning data. The segment number is associated with the partition contained in each of the plurality of quadrants, and the pole pair number represents the pole pair position. The location estimation step has the following characteristics: Step 7, in which the N1 first magnetic sensors are used to acquire the N1 digital signals; Step 8, in which the N2 second magnetic sensors are used to acquire the N2 analog signals; The ninth step is to determine the current quadrant from the plurality of quadrants based on the N1 digital signals obtained in the seventh step. Step 10, which determines the current partition from the plurality of partitions based on the N2 analog signals obtained in step 8; as well as Step 11, based on the learning data, determines the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

2. A position estimation method for estimating the rotational position of an electric motor comprising a rotor having P pole pairs, where P is an integer greater than or equal to 2, the position estimation method being characterized by having: The learning steps involve acquiring learning data needed to infer the rotational position; and The position estimation step, based on the learned data, estimates the rotational position of the rotor. The learning steps have the following characteristics: The first step involves rotating a magnet having a pole pair and sharing a rotation axis with the rotor together with the rotor. In the second step, N3 third magnetic sensors, which are arranged opposite to the magnet and along the rotation direction of the magnet, are used to acquire N3 analog signals whose electrical signals vary according to the magnetic field strength and have a third phase difference with each other, where N3 is an integer greater than 2. The third step utilizes N2 second magnetic sensors arranged opposite to the rotor and along the rotation direction of the rotor to acquire N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other, where N2 is an integer greater than or equal to 3. The fourth step is to calculate the time series data of the mechanical angle during the learning period based on the N3 analog signals obtained during the learning period equivalent to one mechanical angle cycle. The fifth step, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions. as well as Step 6 involves acquiring data representing the correspondence between the time-series data of the mechanical angle and the pole pair number, which is then used as the learning data. The location estimation step has the following characteristics: Step 7, which uses the N3 third magnetic sensors to acquire the N3 analog signals; Step 8, which calculates the current value of the mechanical angle based on the N3 analog signals obtained in step 7; as well as The ninth step, based on the learned data, determines the pole pair number corresponding to the current value of the mechanical angle as the initial position of the rotor.

3. A position estimation method for estimating the rotational position of an electric motor comprising a rotor having P pole pairs, where P is an integer greater than or equal to 2, the position estimation method being characterized by having: The learning steps involve acquiring learning data needed to infer the rotational position; and The position estimation step, based on the learned data, estimates the rotational position of the rotor. The learning steps have the following characteristics: The first step involves rotating a magnet having a pole pair and sharing a rotation axis with the rotor together with the rotor. In the second step, N4 fourth magnetic sensors, which are arranged opposite to the magnet and along the rotation direction of the magnet, are used to acquire N4 analog signals whose electrical signals vary according to the magnetic field strength and have a fourth phase difference with each other, where N4 is an integer greater than or equal to 3. The third step involves using N2 second magnetic sensors arranged opposite to the rotor and along the rotation direction of the rotor to acquire N2 analog signals whose electrical signals vary according to the magnetic field strength and have a second phase difference with each other, where N2 is an integer greater than or equal to 3. The fourth step is to divide the learning period into multiple quadrants based on the N4 analog signals obtained during the learning period, which is equivalent to one mechanical angle cycle. The fifth step, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions. as well as Step 6 involves acquiring data representing the correspondence between the segment number and the pole pair number, as the learning data. The segment number is associated with the partition contained in each of the plurality of quadrants, and the pole pair number represents the pole pair position. The location estimation step has the following characteristics: Step 7, in which the N4 fourth magnetic sensors are used to acquire the N4 analog signals; Step 8, in which the N2 second magnetic sensors are used to acquire the N2 analog signals; Step 9, which determines the current quadrant from the plurality of quadrants based on the N4 analog signals obtained in step 7; Step 10, which determines the current partition from the plurality of partitions based on the N2 analog signals obtained in step 8; as well as Step 11, based on the learning data, determines the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

4. The location estimation method as described in any one of claims 1 to 3, characterized in that, The fifth step of the learning process has the following characteristics: The step of extracting the zero-crossing points, which are the points where the N2 analog signals contained in each of the P pole pair regions intersect with the reference value; The step of extracting the points where the N2 analog signals contained in each of the P pole pair regions intersect each other, i.e., the intersection points; as well as The step of determining the interval between adjacent zero-crossing points and intersection points as the partition.

5. The location estimation method as described in any one of claims 1 to 3, characterized in that, The learning step is performed when the signal processing device is first powered on. The signal processing device performs at least the processing based on the learning step and the position inference step. After the learning step is performed, the position estimation step is performed when the power to the signal processing device is turned back on.

6. A position estimation device for estimating the rotational position of an electric motor comprising a rotor having P pole pairs, where P is an integer greater than or equal to 2, characterized in that the position estimation device comprises: A magnet having a pair of magnetic poles and sharing a rotational axis with the rotor; N1 first magnetic sensors are configured opposite to the magnet and along the rotation direction of the magnet, where N1 is an integer greater than or equal to 3; N2 second magnetic sensors, which are configured opposite to the rotor and along the rotation direction of the rotor, where N2 is an integer greater than or equal to 3; and A signal processing device that processes the output signals of the first magnetic sensor and the second magnetic sensor. The signal processing device has: The processing unit performs learning processing to acquire learning data required to infer the rotational position, and position inference processing to infer the rotational position of the rotor based on the learning data. as well as The storage unit stores the learning data. As part of the learning process, the processing unit performs the following: The first process causes the magnet to rotate together with the rotor; The second process involves acquiring N1 digital signals whose levels reverse and which have a first phase difference with each other when the magnet rotates 180° using the N1 first magnetic sensors. The third process involves acquiring N2 analog signals, each having a second phase difference with the others, through the N2 second magnetic sensors, where the electrical signals vary according to the magnetic field strength. The fourth process, based on the N1 digital signals obtained during a learning period equivalent to one mechanical angle cycle, divides the learning period into multiple quadrants with N1-bit digital values ​​that are different from each other. The fifth process, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions. as well as The sixth process stores data representing the correspondence between the segment number and the pole pair number as the learning data in the storage unit, wherein the segment number is associated with the partition contained in each of the plurality of quadrants, and the pole pair number represents the pole pair position. As part of the location estimation process, the processing unit performs: The seventh process acquires the N1 digital signals through the N1 first magnetic sensors; The eighth process acquires the N2 analog signals through the N2 second magnetic sensors; The ninth process determines the current quadrant from the plurality of quadrants based on the N1 digital signals obtained in the seventh process. The 10th process determines the current partition from the plurality of partitions based on the N2 analog signals obtained in the 8th process; as well as The eleventh process, based on the learning data, determines the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

7. A position estimation device for estimating the rotational position of an electric motor comprising a rotor having P pole pairs, where P is an integer greater than or equal to 2, the position estimation device being characterized in that it comprises: A magnet having a pair of magnetic poles and sharing a rotational axis with the rotor; N3 third magnetic sensors are arranged opposite to the magnet and along the rotation direction of the magnet, where N3 is an integer greater than or equal to 2; N2 second magnetic sensors, which are configured opposite to the rotor and along the rotation direction of the rotor, where N2 is an integer greater than or equal to 3; and A signal processing device that processes the output signals of the second magnetic sensor and the third magnetic sensor. The signal processing device has: The processing unit performs learning processing to acquire learning data required to infer the rotational position, and position inference processing to infer the rotational position of the rotor based on the learning data. as well as The storage unit stores the learning data. As part of the learning process, the processing unit performs the following: The first process causes the magnet to rotate together with the rotor; The second process involves acquiring N3 analog signals, each having a third phase difference with respect to the magnetic field strength, through the N3 third magnetic sensors. The third process involves acquiring N2 analog signals, each having a second phase difference with the others, through the N2 second magnetic sensors, where the electrical signals vary according to the magnetic field strength. The fourth process calculates time-series data of the mechanical angle during the learning period based on the N3 analog signals obtained during a learning period equivalent to one mechanical angle cycle. The fifth process, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions. as well as The sixth process stores the data representing the correspondence between the time-series data of the mechanical angle and the pole pair number as the learning data in the storage unit. As part of the location estimation process, the processing unit performs: The seventh process acquires the N3 analog signals through the N3 third magnetic sensors; The eighth process calculates the current value of the mechanical angle based on the N3 analog signals obtained in the seventh process; as well as The ninth process determines, based on the learning data stored in the storage unit, the pole pair number corresponding to the current value of the mechanical angle, as the initial position of the rotor.

8. A position estimation device for estimating the rotational position of an electric motor comprising a rotor having P pole pairs, where P is an integer greater than or equal to 2, the position estimation device being characterized in that it comprises: A magnet having a pair of magnetic poles and sharing a rotational axis with the rotor; N4 fourth magnetic sensors are arranged opposite to the magnet and along the rotation direction of the magnet, where N4 is an integer greater than or equal to 3; N2 second magnetic sensors, which are configured opposite to the rotor and along the rotation direction of the rotor, where N2 is an integer greater than or equal to 3; and A signal processing device that processes the output signals of the second magnetic sensor and the fourth magnetic sensor. The signal processing device has: The processing unit performs learning processing to acquire learning data required to infer the rotational position, and position inference processing to infer the rotational position of the rotor based on the learning data. as well as The storage unit stores the learning data. As part of the learning process, the processing unit performs the following: The first process causes the magnet to rotate together with the rotor; The second process involves acquiring N4 analog signals, each consisting of an electrical signal that varies according to the magnetic field strength and has a fourth phase difference with the others, through the N4 fourth magnetic sensors. The third process involves acquiring N2 analog signals, each having a second phase difference with the others, through the N2 second magnetic sensors, where the electrical signals vary according to the magnetic field strength. The fourth process divides the learning period into multiple quadrants based on the N4 analog signals obtained during the learning period, which is equivalent to one mechanical angle cycle. The fifth process, based on the N2 analog signals obtained during the learning period, divides the learning period into P pole pair regions associated with pole pair numbers representing the pole pair positions of each of the P pole pairs, further divides each of the P pole pair regions into multiple partitions, and associates a segment number representing the rotation position with each of the multiple partitions. as well as The sixth process stores data representing the correspondence between the segment number and the pole pair number as the learning data in the storage unit, wherein the segment number is associated with the partition contained in each of the plurality of quadrants, and the pole pair number represents the pole pair position. As part of the location estimation process, the processing unit performs: The seventh process acquires the N4 analog signals through the N4 fourth magnetic sensors; The eighth process acquires the N2 analog signals through the N2 second magnetic sensors; The ninth process determines the current quadrant from the plurality of quadrants based on the N4 analog signals obtained in the seventh process; The 10th process determines the current partition from the plurality of partitions based on the N2 analog signals obtained in the 8th process; as well as The eleventh process, based on the learning data, determines the pole pair number corresponding to the segment number associated with the current partition contained in the current quadrant, as the initial position of the rotor.

9. The position estimation device as described in any one of claims 6 to 8, characterized in that, In the fifth process of the learning process, the processing unit performs: The zero-crossing point is the point where the N2 analog signals contained in each of the P pole pair regions intersect with the reference value. The process of extracting the points where the N2 analog signals contained in each of the P pole pair regions intersect each other, i.e., the intersection points; as well as The process of determining the interval between adjacent zero-crossing points and intersection points as the partition.

10. The position estimation device according to any one of claims 6 to 8, characterized in that, The processing unit performs the learning process at least when the power to the signal processing device is first turned on. After performing the learning process, when the power to the signal processing device is turned back on, the processing unit performs the position estimation process.

11. An unmanned transport vehicle, characterized in that, include: An electric motor having a rotor having P pole pairs, where P is an integer greater than or equal to 2; as well as A position estimation device as described in any one of claims 6 to 8 for estimating the rotational position of the electric motor.

12. A sewing device, characterized in that, include: An electric motor having a rotor having P pole pairs, where P is an integer greater than or equal to 2; as well as A position estimation device as described in any one of claims 6 to 8 for estimating the rotational position of the electric motor.

Images