DE19638912A1 - Position sensor arrangement for rotation angle sensor in motor vehicle - Google Patents

Abstract

The sensor includes a coding element (6) which controls a light (L) from a light source (5) hitting on a CCD line sensor, in dependence on the position of the coding element. An evaluation arrangement (V,12,4) determines the position of the coding element in response to a current light distribution on the CCD line sensor (11), whereby the light source is activated in pulses. The light source is preferably driven with the maximum nominal power permissible for pulse operation. The light source is preferably switched on periodically, whereby its activated time is smaller than five percent of the total period.

Description

The invention relates to a position sensor, in particular a rotation angle sensor with a CCD line. Such a sensor has been described in the Texas Instruments Application Brief under the title "linear products TSL 214 integrated opto-sensor" in conjunction with FIGS. 1 to 5 shown there. Details of the integrated opto-sensor TLS 215 were published in the product description by the company Texas Instruments in chapters 4.19 to 4.28 and the heading "TSZ 215, 128x1 integrated opto-sensor" under the number SOES 005A-May, 1993.

The well-known CCD line has a number of side by side arranged collection elements, which are called pixels. These pixels are based on an initial state, in from which the collection elements are emptied, depending on the luminous flux striking them. The on due to the light distribution on the pixels below different charges correspond to different voltages value on the pixels. To a predetermined periodically Repetitive times are the voltages of all pixels to assigned memory elements of a shift register par allel transferred so that the voltage distribution on the individual storage elements in the shift register to the luminous flux supplied to individual associated pixels speaks. Using suitable electronics, the Collecting elements (pixels) on their empty initial state  switched back and by the striking lighting like the charged while at the same time gradually the span voltage values from the storage elements of the shift register read out serially and the analog values into digital values be implemented. So finally, for example 64 or 128 digital values are available, which correspond to the associated pixels during the previous time Describe the luminous flux encountered in the section. With help of a coded aperture having carrier, at which the transmitted light distribution of the position of the Trä gers corresponds to the CCD line, so it is possible Lich, due to the measured light or voltage distribution on the current position of the wearer during measurement close gear. This applies to the application described example of the angle of rotation of a coded disc, however also for the description of the longitudinal position of a beam, if there the linear position of the carrier using a optically transparent coding is described. The use Such optically transparent coding has been used already described in 195 32 903.1.

The essential Un In contrast to the coding disk described, that the luminous flux penetrating the pane is not associated with Hil fe a CCD line, but with the help of one or more Diodes is measured. The invention is therefore based on one Position sensor, especially the angle of rotation sensor the preamble of claim 1 resulting genus. Night egg lig with such a sensor device, it is now that during a sampling period of luminous flux distribution of the luminous flux passing through the code disk, the Code disc with a correspondingly quick turn of the steering wheel significantly shifts. That is, the assignment of individual translucent coding fields on the code disc against above the illuminated pixels changes so that the luminous flux distribution of the incident light during  of a collection period changed. As a result of this fuzzy gray value transitions, so that the image of the light distribution due to the code disc on the CCD line is washed out. To avoid the blurred gray value transitions reduce, one can try the number of code words and thus keeping the coding as large and coarse as possible. A coarse coding also serves to reduce the dirt sensor bility and for simple code disk production.

The before lying invention has improved in known Systems tasked with the gray value transitions other measures greatly improve and so the sensitive speed of the sensor.

The task is characterized by the characteristic part of the combination of features resulting from claim 1. The In principle, the invention therefore consists in the duration of the loading Illumination of the coding element, i.e. at a rotation angle sor of the encoder disc to shorten significantly. So that will be Illuminated coding element only briefly, for example in the form of a strobe. It can be disadvantageous in the invention be that the luminous flux emitted and thus that of the Loads of collecting elements are comparatively small becomes. This can be remedied by the combination of features create according to claim 2. The light source is so driven that reached the nominal power over the entire duration becomes. It is thus expressed differently by the light source emits strong amounts of light for a short time. To now too due to the reduced luminous flux, the position of the coding element after the charging process can, the switch-on time is preferably the light source with the sampling time and the reading timing syn chronized.  

Tests have shown that the cyclical switch-on time ent speaking the features of claim 3 less than 5% of Cycle time and even better less than 1% of the cycle time should be.

In order to measure the accuracy of the sensor the preamble of claim 1 resulting genus still white ter to improve, the invention proposes in a genus according sensor the combination of features according to claim 4. Then the coding element (i.e. at an angle of rotation sor the coding disk) with coding fields, which correspond to only two states. That is, the coding element is provided with coding fields that are either opaque are casual or have a defined translucency. This light transmission corresponds to states 0 and 1 of a multi-step code. Here is the geometric type order made such that always by a coding field several storage elements (pixels) are covered or covered by Luminous flux can be applied. In itself it would also be thought bar to assign a separate coding element to each pixel. Of the The suggestion made here is for greater security to get in Borderline cases to be able to decide whether a sufficient Luminous flux was sent to a group of pixels or not. This makes the decision whether a translucent or impermeable coding field was present, very much safe more than if only one pixel per coding field is provided. It is also the case that several coding fields the (for example 12 coding fields) through the CCD line scanned and evaluated by the sensor, again several pixels are assigned to each coding field, for example between 10 and 11 pixels (this results for example wise when using a 12-step maximum code and the use of a coding line with 128 pixels). The Use of a larger number of pixels per coding element  or coding step of a multi-step code has the before partial effect that there is a light-suppressing coding field or a luminous flux-transmitting coding field even then still shows if it determines the flank of the image taking intermediate voltages. This is for example, the case when the edge of a coding field only partially covers a pixel or if scattered radiation at the edges of the coding field distort the edges of the pixels.

So much for the position sensor according to the angle of rotation determination in particular the determination of the angle of rotation of a Steering wheel is used, is recommended in training the Invention the use of the combination of features according to An Proverb 5. In principle, the coding element can be made from a Coding disk be formed on the similar to cake piece ke following a pie in front of each other through the code written order translucent and light impervious circle sectors arranged adjacent to each other are. You can according to the combination of features Claim 6 the coding fields on a circular ring each other follow, similar to if you made from the pie pieces mentioned only cuts out those sections that are within of the annulus. In this case it is recommended align the CCD line tangent to the annulus and above or below the coding disc above the annulus to arrange. Since the line essentially forms a straight line, while the circular sector is curved, it depends on the permeable coding fields also let through luminous flux on the position of the assigned pixels on the line. Here however, appropriate corrections can be made which compensate for the regular reductions in luminous flux. In further development of the invention, the coding is recommended the combination of features according to claim 6. Then a  Maximum code selected. This maximum code stands out there by from that arranged on an arc in itself orderly sequence of coding elements does not repeat itself encoded values (words lead. In other words can, for example, 180 coding fields on a circular ring be accommodated, always 12 consecutive Code elements result in a twelve-step code word. Will the Circular disk at an angle of 2 °, i.e. around a code element shifted, a new code word results. Every 180 possible Code words are different from each other, always a new first only in the case of adjacent coding fields Step added and the last step omitted (or vice versa with opposite direction of rotation).

To get greater redundancy, you go accordingly the combination of features according to claim 8 by a maximum co de from 4,096 code words, whereby only one is comparative short section of this code word sequence that is used over 180 code words are enough. As such, you are free to choose which one Section with 180 code words is taken from the total sequence of 4,096 code words are cut out and on the code disk creates a circular shape. However, preference is given to a street tegie proceeded, which by the features of claim 9 is described. According to the combination of characteristics according to An Proverb 6 is recommended with a certain design of the coding field, the CCD line tangent to an annulus to arrange.

Another possibility for arranging the coding line is in the combination of features according to claim 10 specified. After that the coding line is arranged radially. This is it necessary that the steps of the code line each read multi-step code word in radial Arrangement to each other. Such a radial arrangement  voltage may be appropriate, for example, if the Comparatively available circumference of the annulus is as small as the surface of a steering column. It but the radially arranged code words do not have to be on the Shell surface of a hollow cylinder can be arranged, they can also run radially on a circular disc. In both cases A Gray code will preferably be used.

To microprocesso when evaluating the code words read To be able to use ren is recommended in further training of Invention the combination of features according to claim 11. After that the evaluated code words are processed by a microprocessor evaluated, for example by checking for errors and based on the code word found, the measured angle is determined.

It has already been explained above that a step of a multi-step code not be by a single pixel is true, but practically the curve contour of this Step described by the voltages of several pixels becomes.

So while knowing the be through the 128 pixels the curve shape matched the corresponding (12-step) Codeword found and so the angle value roughly to 2 ° can be determined, it is recommended to determine the win kels the use of the feature with greater accuracy combinations according to claim 12. After that, not only that multi-step codeword generally determined, but also its position opposite the beginning or end of the CCD line. This results in a significant increase in the Ge accuracy with which the currently measured angle is determined that can. If you find that the edge of a or several steps z. B. by 1 pixel compared to normal  tion to the right or left are shifted thereby the measured angular position with 10 times accuracy Determine accuracy since about 10 pixels per coding step write that results from the coding field. It can you can use different strategies. So at for example the shift of the flanks or the shift Exercise the maximum of one or more code steps a normal position can be determined.

With appropriate evaluation algorithms (software) can be also achieve a resolution in the sub-pixel range. To increase The accuracy of the contrast between the Playback more translucent versus opaque Coding fields can be determined. By polluting the Code fields by partially clogging the translucent Fields with dirt can reduce the difference here values (contrast) can be determined. The contrast drops below a certain threshold, this can mean that the pane must be cleaned or replaced. The Merk Painting combination according to claim 13 also proposes the transmission power of the light source by increasing the transmission adjust performance or extension of the transmission time so that the desired contrast is restored.

This principle applies to all sen working with the CCD line sensors applicable and thus represents another independent Solution in improving sensor sensitivity When evaluating the output from the CCD line Analog values is the respective analog value in one Convert digital number, which is the amplitude value of the Pi xel tension describes and continues due to the digital pay the measured angle of rotation value to be calculated with all the additional calculations such as error correction and others test steps. Because the conversion from analog to digital  value parallel to the calculation of the angular value can, in further development of the invention, the Ver application of the combination of features according to claim 14. This measure can take for all sensors resulting from claim 1 giving genus can be applied. So is a digital value was formed by converting an analog value, so the conversion of the next value using an interrupt prompted and then with the calculation of the angle value continued through the arithmetic unit.

An embodiment of the invention is described below hand of the drawing explained. It shows:

Fig. 1 shows the schematic structure of a CCD line

Fig. 2 in a symbolic representation of the structure of an inventive rotation angle sensor

Fig. 3 shows the structure of a coding disc for a rotation angle sensor with a tangentially aligned CCD line

Fig. 4 shows a section of a maximum code as it is arranged on the disc of FIG. 3

Fig. 5 shows the digitized pixel voltages

Fig. 6, the differentiated voltage values according to FIG. 5 and

Fig. 7 shows the formation of a 12-schrittigen codes where the zero and derived from the individual pixels 1 values are equal, respectively, and the edges of the pulses are determined by the turning points of the slope pulse according to Fig. 5.

FIG. 1 shows the scheme, as described above in connection with the aforementioned references. In Fig. 1, for example, 16 photosensitive pixels 1 are shown , which are acted upon by a luminous flux L. The distribution of the luminous flux striking the individual pixels 1 depends on the coding of a coding element, which in the present example is designed as a coding disk (see FIGS. 2 and 3). The light distribution of the luminous flux L depends on the currently valid code word on the coding disk, which in turn depends on the angle of rotation of the coding disk. Depending on the distribution of the luminous flux L over the individual pixels 1 , a corresponding charge is collected there, which in turn leads to an associated voltage at the output of the pixels. After a predetermined period of time, the gate S is opened via the connections G and U, so that the voltage values of the individual pixels 1 are transmitted in parallel to the associated memory positions 2 of the shift register 3 . This represents an image of the voltage values of the pixels in the associated memory positions of the shift register 3 . Then the pixels 1 z. B. reset to its initial position by discharge and exposed again to the luminous flux L. During this lighting time, the shift register 3 is read out step by step serially to the read-out stage V and is available at the output A ( FIG. 2). After the shift register has been emptied, the voltage values of the pixels 1 which have now formed can in turn be transferred in parallel to the shift register 3 , where they are then read out serially.

Fig. 2 shows the operation of the sensor in a symbolic representation. The infrared LED as the light source 5 , which is pulsed as a function of the computer 4, emits a light band onto the coding element 6 , which is designed as a coding disk.

The coding disk has sector-shaped translucent fields of FIG. 7 and opaque fields 8 , 12 of these adjacent fields each giving a 12-step code word. The code on the coding disc 6 is a maximum code, ie the individual code words are created by a window having 12 steps. If the code disk 6 rotates by one step, a new code word results from the fact that a new first step has been added and the last step of the previous word has been omitted. None of the possible code words on the disc are repeated. The coding fields 7 , 8 are either non-permeable or non-permeable, with several coding fields having the same permeability lying next to one another and thus resulting in a correspondingly large sector of a circle. The planar light bundle 9 strikes the coding disk 6 and who passes the parts of the light bundle through corresponding coding fields 7 , while other parts of the bundle are not passed to the CCD line 10 . This results in a light distribution L on the CCD line 10 , which strikes the pixel row 11 and charges the pixels accordingly. The pixel voltages are then, as already described in connection with FIG. 1, periodically reloaded onto a shift register and then fed serially via a read-out stage V to an analog / digital converter 12 . This analog / digital converter can still be integrated in the CCD line 10 , which results in a so-called integrated opto-sensor. It is also possible to use the internal micro-controller A / D converter.

The digitized output 13 of the converter 12 is fed to the computer 4, which in turn 5 controls the pulse-wise operation of the LED and the resetting of the pixel line 11 and the withdrawal from the not shown in detail shift register via control lines 13,14. From the digital values at the output 13 , the corresponding code word is then finally recognized in the computer 4 and the measured angle of rotation of the coding disk 6 is roughly determined therefrom. Fine winding determination by evaluating the edges (0-1 transitions) or centers of gravity of the 0 or 1 fields.

Fig. 3 shows the coding disk 6 with the individual coding fields 7 , 8 , which correspond to 0 and 1 values. On the side opposite the light source 5 , as already indicated in FIG. 2, the CCD line 10 is arranged in the tangential direction and in such a way that a 12-step code can be read by it. Assuming that the CCD line 10 is provided with 128 pixels and that 180 different angular values can be described as a rough measurement using the 12-step code, each step of the code word corresponds to approximately 10 pixels. The maximum code used is indicated in FIG. 3, in which the individual steps, which in turn correspond to coding fields 7 , 8, were designated 0 or 1 (0 corresponds to impermeable, 1 corresponds to permeable). If the disc rotates below the CCD line 10 , a new step appears above the start of the line, while at the end of the line the last step has no effect on the line.

The effect of the maximum code is indicated in FIG. 4. Here, the result of the code fields 7, 8 during rotation of the encoder disc 6 under the pixel line 11 to the right or left depending on the direction of rotation of the encoder 6 pushes. In the drawing, this is shown as a shift of the pixel window to the right or left into position 11 a or 11 b. This results in corresponding new 12-step code words.

FIG. 5 shows the digitized voltage values on the output line 13 in FIG. 2, which correspond to the individual 128 pixels. An envelope curve is clearly shown, from which the code word currently read can be determined, which in turn describes the angle of rotation of the coding disk 6 that has just been measured. In order to recover the code word, the adjacent voltage values are differentiated ( FIG. 6) and the extreme values of the differentiation curve according to FIG. 6 form the edge positions for the processed code word. In Fig. 7, the respective digital value of the 12-step code word was written under the transformed pixel voltages. The code word specifies the angular position of the coding disk 6 exactly to 20. However, it is possible to determine the measured angle much more precisely by calculating the shift of the code word on the axis with respect to a zero position. As can be seen from FIG. 6, the code word 7 is shifted a little to the right, since a single coding field 7 , 8 corresponds approximately to 10 pixels. From the shift of the code word in FIG. 7 with respect to the zero position, it can be specified whether the actual angle of the coding disk is now somewhat larger or slightly smaller than the previously roughly determined 2 ° value. Possible resolutions of approximately 0.1 ° can be achieved by taking into account the shift in the code word relative to the CCD line 10 . It is also important for the invention that the maximum code used is only a section of a code word comprising 4,096, the selected section comprising 180 code words. This results in a high probability of recognizing errors without taking other information into account, namely a probability of 96% (W = 4,096-180: 4,096 = 0.96). A further optimization of the code up to the 100% probability of error detection is possible by considering other criteria such as B. Minimum Hamming distance.

Overview introduction

Digital sensors are often used for steering wheel angle determination discrete optical or magnetic construction elements used. The achievable resolution is essential Lichen from the number and the positioning accuracy of the used sensors, the mechanical boundary conditions as well depending on the tolerance situation and often not enough reaching. To increase the resolution, one can be described above benes optical principle used on a CCD (Charge-Coupled-Device) line based.

A CCD line is an optical sensor sor, which consists of a large number (e.g. 64, 128, 256, 1024,...), sensor elements arranged linearly in a housing consists. The sensor corresponding to the gray value information voltages and currents can be already integrated electronics can be read out. As a result of high pixel count is a processing of the records in the Mi croprocessor according to digital image processing possible where by u. a. a very fine angle determination can be achieved can.

Operating principle

The entire sensor essentially consists from a light source, a coded disc (or similar), the CCD line and the evaluation electronics. From the code disks conditional gray value distribution on the optosensitive area The line determines the steering wheel angle position.

When using a code disk, the CCD line can both be positioned tangentially as well as radially to this. The code (e.g. Ma  ximalcode) on the circumference of the code disk, with radial mounting the code (e.g. Graycode) is circular in the disc to integrate. As an alternative to using a code disk can encode other components to reduce space ten (e.g. steering column; axial and tangential sensor position possible). Depending on the boundary conditions construction, pollution problems tic, etc. is the principle of transmitted light or reflection to prefer.

For easy manufacture and improvement of dirt Sensitivity is "rough" code structures (broad te 0-1 info) to use. When using a code disk with tangential sensor position, the absolute angle can be over the actual code can be roughly determined. A fine win Kel resolution is achieved by z. B. the edges of the 0-1 transitions or the focus of the light and dark fields be determined and evaluated.

The sampling time of the coding with the CCD line is not shorten as you like. The result is faster steering wheel rotation on the line resulting in unsharp gray value corridors can be improved by not using the light source in continuous operation but pulsed is used. The one switching times are with the sampling and readout time ming to synchronize (e.g. control of the light source with the micro controller).

With suitable software algorithms, tolerances (e.g. Positioning accuracy of the sensor, radial and wobble movement of the code disc, aging of the light source,. . .) not available or can be compensated within certain limits. To Error detection mechanisms can also be integrated.  

The figure shows a possible concept of evaluation electronics tronics. The gray value information from opti sensors of the CCD line, controlled by the micro-con troller, sequentially off, with an A / D converter digitali and then read into the controller. At An external A / D converter is required for fast conversion times be, at lower clock frequencies the in is sufficient converter integrated into the controller. The one after the selection process is completely available in the microcontroller The existing data record is then created with the corresponding software re evaluated. Is the processing of the gray value information completed, the reading of the next data set begins zes.

In general, the use of a CCD line is not only in the Steering angle sensor application conceivable. It can be general for linear and rotary position detection used will.

As mentioned, is used to minimize the need for digitization a fast external A / D converter (ADU) is required (is the ADU integrated in the microcontroller relatively slow). To save the expensive external The following procedure is possible for ADUs: During the loading Data record already read in the previous cycle is evaluated the individual analog values of the next th record in the internal microcontroller A / D converter "slowly" digitized and saved. I.e. after a Digitization process is completed by a Interrupt the evaluation routine of the data record interrupted, the digitized value now available is saved saved, the conversion of the next analog value started and continue with the previously canceled routine. this will  repeated until all analog values of a data set are digitized.

Claims (16)

1. Position sensor, in particular rotation angle sensor for steering angle measurement of motor vehicles, in which the luminous flux (L) of a light source ( 5 ) incident on a CCD line is controlled by a coding element ( 6 ) as a function of the position of the coding element and in which an evaluation device (V, 12.4 ) determines the position of the coding element ( 6 ) on the basis of the instantaneous light distribution on the CCD line ( 11 ), characterized in that the light source ( 5 ) is switched on in pulses. 2. Position sensor according to claim 1, characterized in that the light source ( 5 ) is operated at the maximum rated power for pulse operation. 3. Position sensor according to claim 1 or 2, characterized in that the light source ( 5 ) is switched on cyclically and the switch-on time is less than 5%, in particular 1% of the cycle time. 4. Position sensor in particular rotation angle sensor for steering angle measurement of motor vehicles, in which the incident on a CCD line ( 11 ) incident light flux (L) of a light source ( 5 ) by a coding element ( 6 ) depending on the position of the coding element ( 6 ) controls ge and in which the position of the coding element ( 6 ) is determined by an evaluation device (V, 12 , 4 ) based on the instantaneous light distribution on the CCD line ( 11 ), characterized in that the coding element ( 6 ) with Coding fields ( 7 , 8 ) is provided, which are essentially translucent or opaque, with several such coding fields ( 7 , 8 ) depending on the position of the coding element ( 6 ) the distribution of the light source ( 5 ) to the CCD Determine line ( 11 ) of luminous flux (L) falling on individual pixels ( 1 ). 5. Position sensor according to one of the preceding claims, characterized in that the coding element ( 6 ) is a coding disk and the coding fields ( 7 , 8 ) are preferably substantially by adjacent circle sectors with the same sector angle or sections of these sectors. 6. Position sensor according to claim 5, characterized in that the coding fields ( 6 ) are arranged in succession on a closed ring, which is preferably located on a circular disc and the coding line ( 11 ) is aligned tangentially to the ring. 7. Position sensor according to one of the preceding claims, characterized in that the coding element has the shape of a circular disk ( 6 ) or a lateral surface of a cylinder and that the coding fields ( 7, 8) are arranged in succession on several concentric circular rings and that CCD line ( 11 ) is arranged such that it extends over several circular rings, the Co dier values scanned by the CCD line preferably corresponding to a Gray code. 8. Position sensor according to one of the preceding claims, characterized in that the light-permeable and opaque coding fields ( 6 ) are arranged in the form of a multi-step, for example 12-step maximum code. 9. Position sensor according to claim 8, characterized net that the maximum code used from a sequence from Z. B. 180 code words with 12 steps each stands, which consists of a longer maximum code with z. B. a move of 4,096 code words the same step number are cut out. 10. Position sensor according to claim 8, characterized in that the code word sequence from the code word train are selected according to the following sequence of specifications:

• a) the code word sequence must again result in a self-contained maximum code,
• b) as short as possible sequences of identical translucent or opaque code fields should appear,
• c) there should be as many changes as possible between translucent and opaque code fields,
• d) There should be as large a Hamming spacing as possible from neighboring code words.

11. Position sensor according to one of the preceding claims, characterized in that the evaluation device (V, 12 , 4 ) converts the voltage pulses emitted by the CCD line ( 11 ) into corresponding digital values. 12. Position sensor according to claim 11, characterized net that the evaluation device by the sequence curve curve depicted by digital values after the Descriptive features such as curve flanks and / or extreme values and then the rela tive location of the curve in relation to an egg  ner normal position determined (rough angle determination by determined code word). 13. Position sensor according to one of the preceding claims, characterized in that the evaluation device (V, 12 , 4 ) is provided with a control device which, when the maximum amplitudes fall, values of the measured voltages on the pixel line or the corresponding digital values adaptively the transmit power or the light intensity of the light source ( 5 ) increases. 14. Position sensor according to one of the preceding claims, characterized in that the evaluation device is provided with a microcontroller ( 4 ), that the microcontroller has an arithmetic logic unit and an AD converter which work in parallel in time and that after the conversion of an analog value In a digital value, the calculation of the curve profile and its evaluation are interrupted by an interrupt, a command to convert the subsequent analog value is issued until all pixel voltages are digitized and the calculation of the previous data set is then completed. 15. Position sensor according to one of the preceding claims, characterized in that the coding element ( 6 ) is arranged on the steering column of a motor vehicle and the coding is formed according to the reflection principle. 16. Position sensor according to one of the preceding claims che, characterized in that optionally the pressure light principle or the reflected light principle applied becomes.