WO1999046612A1

WO1999046612A1 - Optical sensor system for detecting the position of an object

Info

Publication number: WO1999046612A1
Application number: PCT/DE1999/000620
Authority: WO
Inventors: Wendelin Feiten; Laszlo Domjan; Janos Giber; Laszlo Kocsanyi; Peter Richter; Gabor Szarvas; Sandor Varkonyi
Original assignee: Siemens Aktiengesellschaft
Priority date: 1998-03-10
Filing date: 1999-03-09
Publication date: 1999-09-16
Also published as: EP1062524A1; CA2322419A1; KR20010041694A; CN1292878A; JP2002506976A

Abstract

The invention relates to a special light source which produces a horizontal strip of light. Said strip of light is reflected by objects in the vicinity of the sensor system and conducted to a photoelectric converter by a special imaging device. Said imaging device is configured in such a way that objects situated further away are represented as being somewhat further apart so that an ordinary objective with linear resolution can obtain better position resolution of objects situated further away from the sensor system over the entire imaging area. Advantageously, light-emitting diodes are provided on the optical axis of a cylindrical mirror as the light-emitting elements. The electrical signals emitted by the photoelectric converter are evaluated by an evaluation unit in relation to their position. Using triangulation, said evaluation unit then determines how far away the objects which reflected the light are situated.

Description

1 description

Optical sensor system for determining the position of an object.

The invention relates to an optical sensor system for exploring objects and for determining their positions. This sensor system can preferably be arranged on autonomous mobile systems in order to enable them to find their way in an unknown environment.

Autonomous mobile systems in development and planning will increasingly be found in household environments in the future. They will carry out transport and cleaning tasks there by autonomously carrying out the tasks imposed on them with the help of an orientation system that enables them to get an idea of their surroundings. In addition to the different route planning and evaluation algorithms, the sensor system is of great importance for the detection of obstacles in the vicinity of the autonomous mobile unit. In order to make autonomous mobile systems attractive to consumers, it is particularly important that they can be mass-produced inexpensively and in a technically simple manner. The sensor systems for detecting the position of objects in the vicinity of the autonomous mobile systems must therefore be robust and inexpensive to manufacture.

Optical sensor systems which perform the detection of objects by triangulation are known from the prior art. A special measurement method is, for example, to carry out an active optical triangulation with strip lighting. The main elements of such a measuring system consist, for example, of a light source that illuminates a strip in space, an optical imaging system, a two-dimensional image receiver and electronics for processing and evaluating the images 2 signals received by the receiver. In order to carry out the triangulation, such a system requires a light source that only illuminates a strip of space. An important feature of this light source is the surface density of its emitted light output. The requirements for such light sources are that the power density must be large enough to be able to recognize even dark objects. Another feature of such light sources is the thickness of the illuminated strip, which influences both the size of the detectable spatial area and the resolution when measuring the position of detected objects.

The following options are currently known for the implementation of light sources for the emission of light strips: a cylindrical lens in front of a collimator lens which collimates the light from a laser source or an incandescent, halogen or arc lamp; the parallelization of light from an incandescent or halogen lamp with the aid of a collimator lens and its fanning out by a cone mirror. The imaging system used must above all meet two main requirements. On the one hand, it has to map the spatial area to be measured onto the surface of the two-dimensional image detector; on the other hand, it must ensure the desired position resolution with regard to the objects to be measured. In this context, the position resolution capability is understood to be the smallest distance between objects, which can still be resolved after being imaged on the optoelectric converter. It is determined by the transducer used as well as by the imaging system. The task of imaging a large spatial area, as far as possible in the form of an entire half-space, and thereby generating a sufficient range resolution capability with the same lens, are contradictory requirements. One possible means of enlarging the measurable spatial area by means of such an imaging system is to use an imaging system Wide-angle lens that creates a distortion-free image. Such lenses with multiple lenses have that 3 disadvantage that they cannot map the entire half space around the lens. Other lenses that represent an entire half-space no longer work without distortion and are expensive to purchase. Distorting lenses also influence the position resolution during triangulation. The distance in the image plane between the images of the backscattered light from two neighboring objects that are close to the imaging system is shown larger than with two neighboring objects that are further away from the imaging system. This leads to the position resolution becoming increasingly poorer with increasing distance from the lens or imaging system. By using normal or aspherical refractive surfaces, such as are common with normal lenses, objects that are far away cannot be imaged sufficiently from one another by the imaging system. The imaging of such objects would require the use of an optical imaging element optimized for this special purpose, which is not known from the literature, however.

From the European publication EP 0 358 628 A2 with the title "Visual navigation and obstacle avoidance structured light system", a triangulation system for use on mobile vehicles is known, which works in an imaging system by means of strip lighting and a normal lens. The disadvantages This solution consists in that on the one hand the vehicle can only detect objects lying in the front, in the direction of travel and on the other hand only a limited area or nearby objects with sufficient resolution can be used for the distance measurement of objects by using a normal lens RA Jarvis, JC Byrne: "An automated guided vehicle with map building and path finding capabilities"; in RC Bolles and B. Roth, editors, 4th International Symposium on Robotics Research, p. 497-504, MIT Press, Cambridge, Massachusetts, 1988, and Y. Yagi, Y. Nishizawa, M. Yachida: “Map based navigation of the mobile robot using 4 omnidirectional image sensor COPIS, Proc. From the 1992 IEEE International Conference on Robotics and Automation, Nice, France, May 1992, it is known to carry out the imaging of the surroundings by means of a cone mirror and a lens. The cone mirror used does not change the resolving power of the system with lenses located far away . It is still determined by the camera lens. In the article by J. Hong, X. Tan, B. Pinette, R. Weiss, EM Riseman; "Image-based homing, Proc. From the 1991 IEEE International Conference on Robotics and Automation, Sacramento, California, April 1991 it is known to map the environment using a spherical sphere. However, no strip lighting that could serve for triangulation is used. Such so-called passive systems have the disadvantage that the objects are not illuminated by a light beam with a known altitude and position information is therefore very difficult to derive since triangulation cannot be carried out. Real-time image processing systems, which require a large amount of computing capacity, must be used for the evaluation. Furthermore, from the article by P. Greguss: "PAL optics-based instruments for space research and robot technology, in lasers and optoelectronics 28 (5) / 1996, pages 43-49 a PAL lens for the use of navigation tasks at au is known - use tonomobile mobile robots. The PAL lens from

Greguss is a wide-angled imaging element that contains two reflective and one refractive aspherical surface and that is capable of imaging an entire half-space. The application of PAL optics as an imaging element of an active triangulating obstacle detection system for robots is also described there.

The invention has for its object to provide an optical sensor system that on mobile vehicles such. B. autonomous mobile robots can be used, which is technically simple and which enables detection of obstacles in all directions around the vehicle, whereby 5 whose imaging system has both the position of objects in the vicinity of the system at distances of less than 50 cm, and in the far area of the system at distances of more than 2 m, a sufficient position resolution of approximately 5-10 cm and an angular resolution of <1 °.

This object is achieved according to the features of patent claim 1. Further developments of the invention result from the dependent claims.

The sensor system described advantageously consists of light sources which illuminate the surroundings of the autonomous mobile unit in the form of strips around the unit, since this enables simultaneous detection of obstacles or objects around the unit. A plurality of light sources arranged one above the other can advantageously also be provided, which are switched on at different intervals, so that different height dimensions of the room can be detected or measured. The light, which is backscattered by objects that are illuminated, is advantageously carried out by using a special wide-angle imaging element that has only a single curved, spherical or aspherical, reflective surface for guiding the light in connection with an objective and a filter, as well as a photoelectric converter , the environment being projected onto the transducer by the imaging system. With this arrangement, the task can be solved with as little technical effort as possible. In a further development of the invention, the best spatial coverage and the best position resolution is advantageously achieved by describing the shape of the aspherical reflecting surface of the wide-angle imaging element with the aid of spline functions. The spline function describes the shape of the imaging element in the following way: the spline function expands areas that are further away depending on the type of lens used. Using the spline function, the 6 distance ranges and the partial areas of the aspherical imaging element that are valid for the respective distance areas are described in such a way that the neighboring polynomial functions in the respective transition points of the detection areas have the same value and the same derivatives, whereby the function used becomes continuous and unbreakable. By using such an imaging element, it is advantageously achieved that simple lenses with a normal viewing angle can be used for the wide-angle linear imaging characteristic required in the development of the invention. Through the use of spline functions it can be achieved that light which is scattered back from more distant areas is deliberately pre-distorted before it passes through the lens, so that more distant areas can be displayed with a higher resolution than it would normally be the lens used would be possible. In this way, an imaging system is made available that offers a simple economical solution for the production of wide-angle, linear optical systems.

Exemplary embodiments of the invention are explained in more detail below with the aid of illustrations.

Figure 1 shows an embodiment of a sensor system. Figure 2 shows a possible arrangement of a light source used in side view. Figure 3 shows a possible arrangement for generating a light streak around the sensor system. Figure 4 shows a further embodiment for producing a light strip around the sensor system. Figure 5 shows a further embodiment for producing a light strip around the imaging device. 7 Figure 6 shows a possible installation of the sensor system on a vehicle or robot in a side view. Figure 7 shows a top view of the sensor system and a vehicle.

Figure 8 shows a sensor system for use with a one-dimensional photoelectric converter. Figure 9 shows an embodiment with a two-dimensional optical position detector.

As Figure 1 shows, there is a possible embodiment of a sensor system described, which illuminates a room strip all around from 4 light sources 1. When arranging these light sources for illuminating a light strip, it should be noted in particular that these light strips lie in a plane which is essentially plane-parallel to the The surface on which the autonomous mobile unit to which the sensor is attached moves. If this plane-parallel arrangement is not possible, the triangulation is made more difficult by taking into account when evaluating the reflected light beams that they have hit the reflecting objects at different angular positions, so that the triangulation is used to determine the distance of the objects result in different triangulation angles. The light sources 1 generate light strips 2 that illuminate the room. The imaging element 4 used projects the light backscattered from the objects 3 through the lens 6 onto the photoelectric converter 7, which in this arrangement is designed as a two-dimensional CCD image detector of a camera 5. The photoelectric converter 7 is connected to an evaluation electronics 8, for example in a computer, which is part of the sensor system for determining object positions. The evaluation electronics 8 determine the position of objects 3 based on the image projected on the photoelectric converter 7 using the principle of active optical 8 triangulation, the imaging properties of the objective 6 and the imaging element 4 in particular being used in conjunction with the height of the plane at which the light strip is emitted. The sensor system is advantageous on mobile vehicles 12, such as. B. used a mobile robot. The current information about the control of the vehicle 12 or the robot can be determined by the evaluation electronics 8 and, based on this information, the further travel path of the unit can be planned. This information gives z. B. the positions of objects 3, under which the vehicle 12 moves through, or between which the vehicle 12 must move.

In the embodiment shown in FIG. 1, the optical axis of the imaging element 4 is advantageously aligned perpendicular to the light strip 2, which are emitted by the light sources 1. In this embodiment, the imaging element 4 is designed as such an optical element which has a spherical or aspherical reflecting surface 9, the outside of the reflecting surface 9 being used for imaging the reflected light beams in the imaging system. The light strips emitted by the light sources and light beams reflected by the objects 3 are projected via the imaging element 4 through the lens 3 onto the light detector 7, as is shown schematically by the numbered beam paths. The distance of the light rays incident on the photodetector 7 is dependent on the imaging properties of the objective and the imaging element 4 and is characteristic of the distance of the objects 3 from the sensor system. In the embodiment of the sensor system shown in FIG. 1, any two-dimensional image detector 7 can be used. For example, a photodiode matrix can be used as a photoelectric converter 7 instead of a two-dimensional CCD sensor. 9 Figure two shows a possible basic embodiment of the applied light source 1 in a side view. The light source 1 shown in Figure 2 consists of a cylindrical mirror 11, which suitably has an aspherical cross section and consists of light emitters 10, which can be light-emitting diodes, for example. These light emitters are located in the focus line of the cylinder mirror 11. The light emitters 10 are arranged one after the other in the focus line of the aspherical cylinder mirror 11, so that the light-emitting surfaces of the light emitters 10 point in the direction of the aspherical cylinder mirror 11. The light emitted by the light emitters 10 first reaches the aspherical cylindrical mirror 11 and then emerges as a strip of light 2 projected through the mirror. By using an aspherical mirror shape adapted to the light emitters 10, it can be achieved that the light source 1 illuminates a parallelized light strip. In this context it should be mentioned that light-emitting diodes (LED) represent an inexpensive, simple and small light source and that due to the selected arrangement of the

LEDs on the axis of the cylinder mirror an inexpensive light source can be provided.

As Figure 3 shows, there is a possible basic arrangement for generating a light strip 2 from part of a cylinder mirror. Here, the cylinder mirror 11 is in a so-called off-axis arrangement with the light emitters 10, which is why only part of the cylinder mirror 11 previously described in FIG. 2 is used. The light emitters 10 are advantageously designed as LEDs and are arranged in succession on the focal line of the aspherical cylinder mirror in such a way that their light-emitting surfaces are directed in the direction of the aspherical cylinder mirror 11. With such a light source, the parallelized light beam valid for the sensor system can also be generated. This is indicated by the beam path provided with the arrows. 1 0

As shown in Figure 4, another possible solution for producing a light strip 2 which shines all around the imaging device is that a parallelized light beam is set in rotation. The rotating light beam 2 is generated in such a way that both the light emitter 10 and the collimator optics 14 are set in rotation about an axis t. A further embodiment of this type can consist, for example, in that both the light emitter 10 and the collimator optics 14 are fixed and the light steel 2 is moved all around with the aid of a rotating mirror.

Figure 5 shows a further possible embodiment for producing a light strip 2 emitted all around. In this embodiment, the light strip 2 is generated by a conical mirror 15. The light emitted by the light emitter 10 is first parallelized by collimator optics 14 and then fanned out into the desired light strip 2 by a conical mirror 15. In this embodiment, for example, an incandescent lamp, halogen lamp, arc lamp or laser can be used as the light emitter 10.

Figure 6 shows the possible structure of a sensor system on an autonomous mobile unit 12, which can be, for example, a service robot. Figure 6 shows a side view. A plurality of light strips arranged one above the other, which are emitted by a plurality of light sources 1 lying one above the other, can advantageously be generated on mobile systems. The light strips 2 are preferably generated at different times and illuminated in a pulsed manner. By arranging several light sources one above the other, a better height differentiation of the obstacles is achieved. In order to be able to detect and measure as many obstacles as possible around the mobile vehicle 12, two imaging elements 4 are preferably used with the cameras 5 associated with them 11 provided at two opposite corners of the mobile vehicle 12. The evaluation electronics ensure that when triangulating obstacles, the corresponding altitude of the currently switched on light source is used to evaluate the triangulation results.

Figure 7 shows the top view of a mobile vehicle 12 provided with the sensor system, e.g. B. A robot with the reception area of the detection system 13. As FIG. 7 further shows, the individual imaging elements 4 are attached to 2 opposite corners of the mobile system 12. If two imaging elements 4 are arranged in the manner shown in Figure 7, the detection area 13 of the optical sensor system can be extended to the entire environment of the mobile system 12 or to the entire space surrounding the mobile system.

Figure 8 shows a possible embodiment of the photoelectric converter 7 in the form of a one-dimensional photoelectric converter 7. The light from the imaging element 4 is projected through the lens 6 onto the one-dimensional light detector, which moves in the image plane or is advantageously rotated. This photoelectric converter 7 can be used, for example, as a one-dimensional position-sensitive detector, e.g. B. as a CCD or PSD. Because the one-dimensional light detector is moved in the image plane or is advantageously rotated, it detects the light intensity distribution in the entire image plane, whereupon the imaging element 4 and the lens 6 image the area around the mobile vehicle 12. The measurement results obtained in this way can preferably be temporarily stored or the evaluation is carried out synchronously with the speed of the photoelectric sensor.

Figure 9 shows another possible embodiment of a photoelectric converter 7, which here is a two-dimensional positive 12 position-sensitive detector is shown. In this embodiment, the photoelectric converter 7 is designed as a two-dimensional position-sensitive detector, which is located behind the lens 6 in its image plane. Between the ob ective 6 and the photoelectric converter 7 there is an impermeable disk 16 in this embodiment, which is provided with a gap 17. When this disk rotates, the area above the position-sensitive detector that is currently swept by the gap 17 is always released. For example, the gap 17 on the opaque pane 16 only allows the light from a well-defined spatial area to pass through, for example with the opening angle of 1 °. In this way, a directional resolution capability with an arbitrarily small angle can be achieved. The slit width to be selected depends on how much light is reflected back, or with what sensitivity the detector works and with what intensity the light strip is illuminated by the light source. When using a rotating parallelized light beam as a light strip 2, as shown in the exemplary embodiment of Figure 4, the use of the disk 16 is not necessary because the directional resolution is already ensured by the rotating light source. All in all, the advantage of the sensor system described is that its reception range is larger than in known embodiments and that the image is more uniform than in other known triangulating sensor systems. The wide-angle image enables the positions of objects that are far from the sensor system to be measured. The special shape of the imaging element 4 ensures the uniform resolution of the distance measurement in the entire detection range of the sensor system by virtually correcting the lack of resolution of the lens 6 in the distance of the sensor system, because it scatters reflected light rays and thus pulls apart more distant objects .

Claims

13 claims

1.Optical sensor system for detecting the position of an object, with a light source for illuminating the environment, a photoelectric converter which converts the intensity distribution of the light scattered back from objects into electrical signals and passes it on to a sensor signal evaluation system, characterized in that the sensor system has a light source ( 1), whereby one or more light strips are generated in the horizontal direction in several spatial directions, and which has an optical imaging element (4) in order to reflect the light scattered back by objects (3) and to project it onto the photoelectric converter (7), the optical imaging element (4) having only a single reflecting surface (9).

2. Optical sensor system according to claim 1, which has a lens (6) in the beam path between the imaging element (4) and the photoelectric converter (7).

3. Optical sensor system according to one of claims 1 or 2, in which the optical axis of the imaging element (4) is perpendicular to the illuminated light strip (2) and has four light sources (1).

4. Optical sensor system according to one of the preceding claims, which has a wide-angle imaging element (4) which has only a spherical or aspherical reflecting surface (9).

5. Optical sensor system according to one of the preceding claims, 14

in which the imaging element (4) has at least one aspherical reflecting surface (9) that can be described by two spline functions.

6. Optical sensor system according to one of the preceding claims, which has at least 2 superimposed light sources (1) in order to produce two superimposed light strips (2).

7. Optical sensor system according to one of the preceding claims, which has as a photoelectric converter (7) a one-dimensional position-sensitive light detector which is moved.

8. Optical sensor system according to one of claims 1-6, which has a two-dimensional position-sensitive detector as a photoelectric converter (7), a spatial light modulator being mounted in the beam path between the imaging element (4) and the photoelectric converter (7).

9. Optical sensor system according to claim 8, in which a rotating disk (16) provided with a gap, or a liquid crystal modulator is provided as the spatial light modulator.

10. Optical sensor system according to one of claims 1-6, wherein the photoelectric converter (7) is designed as a two-dimensional image detector matrix.

11. Optical sensor system according to one of claims 1-5, wherein the light source (1) is designed as a rotating light source. 15

12. Optical sensor system according to any one of claims 1-10, wherein the light source (1) from light-emitting elements (10) is constructed and consists of a cylindrical reflecting surface (11) with an aspherical cross section, the light-emitting elements (10) as Light-emitting diodes are placed in the focus line of the cylindrical mirror surface (11) so that the light-emitting surfaces of the elements (10) face the cylindrical mirror (11).

13. Optical sensor system according to one of claims 1-10, wherein the light source (1) for generating the light strip

(2) from a known light-emitting element, such as an incandescent lamp, a halogen lamp, an arc lamp or a laser, from a collimator lens (14) and from a cone mirror (15).

14. Autonomous mobile unit with a sensor system according to one of claims 1 to 13