SE536983C2 - Control system and method for controlling vehicles when detecting obstacles - Google Patents

Abstract

A control system comprising one or more sensing units, each of which is adapted to sense the oncoming road path for one of the wheels of the vehicle. By identifying particularly dangerous objects as so-called "cat skulls" in this defined roadway, the vehicle can steer on the side and avoid these objects. A "cat's head" mining environment is a fist-sized rock with sharp edges, and is known for puncturing the vehicle's wheels. By using the control system to identify such objects, the vehicle can avoid having its wheels punctured and thus avoid interruptions in production because the vehicle needs to stand still and block the production line, or deviate from its work to have its wheels repaired. (Figure 2)

Description

536 983 carry out their task as quickly as possible without making any mistakes. Autonomous vehicles have, among other things, been developed to be used in dangerous environments, for example in the defense and war industry and in the mining industry, both above ground and underground. If people or ordinary, manually controlled vehicles approach the working area of the autonomous vehicles, they normally cause a break in work due to safety reasons. When the work area is free again, the autonomous vehicles can be ordered to resume work.

The autonomous vehicle uses information regarding the road, the surroundings and other aspects that affect the travel to automatically regulate the throttle, braking and steering. A careful assessment and identification of the planned progress is necessary to assess whether a road is passable and is necessary to be able to successfully replace a person's assessment when it comes to driving the vehicle. Road conditions can be complex and when driving a normal driver-controlled vehicle, the driver makes hundreds of observations per minute and adjusts the operation of the vehicle based on the perceived road conditions to find, for example, a passable road past objects that may be on the road. In order to be able to replace human perception with an autonomous system, this means, among other things, being able to perceive objects in an accurate way in order to be able to effectively regulate the vehicle so that you steer past these objects.

The technical methods used to identify an object in connection with the vehicle include using one or more cameras and radar to create images of the surroundings. Laser technology is also used, both scanning lasers and fixed lasers, to detect objects and measure distances. These are often referred to as LIDAR (Light Detection and Flanging) or LADAR (Laser Detection and Ranging). In addition, the vehicle is equipped with various sensors, among other things, to sense speed and accelerations in different directions. Positioning systems and other wireless technology can also be used to determine if the vehicle is approaching, for example, an intersection, a narrowing of the road, and / or other vehicles. 10 15 20 25 536 983 Autonomous vehicles are used today as load carriers in, for example, the mining industry - both in opencast mines and underground mines. A vehicle breakdown in a bottleneck such as a transportied or in a mining town in many cases immediately stops the entire production chain with significant loss of income as a result. A common reason for vehicle breakdowns in off-road environments is punctures caused by sharp edges on fist-sized stones which in the mining industry are called "cat skulls". The driver of a manually controlled vehicle has the task of seeing and not driving on these cat skulls with any of the vehicle's wheels. For an autonomous vehicle, it is a great challenge to detect these objects because they are relatively small and have an appearance that does not differ much from the substrate in mines.

Another problem with autonomous vehicles is that it can be difficult to know where they plan to drive, unlike vehicles that have a human driver looking at the slope they intend to drive.

US-6151539-A describes a system for autonomous vehicles and a method for controlling them. The system consists of a series of sensors that will ensure that the vehicle is on its way and ensure that the vehicle avoids collision with various obstacles. A downward-facing laser scanner can be used to detect holes or small obstacles on the ground.

US-20090088916-A1 describes a system for autonomous vehicles that should be able to automatically plan their way and at the same time avoid colliding with different types of obstacles. The system uses mathematical algorithms to calculate the correct path and how to avoid obstacles. The system uses various sensors, including lasers, to capture the necessary information and data.

US-20120035788-A1 describes a system for autonomous vehicles which is to detect various obstacles on the road, and which uses a laser to calculate the best route. 20 25 30 536 983 US-20100114416-A1 describes a system for autonomous vehicles that uses a laser scanner that scans the surroundings to get information about what the road looks like and if there are obstacles on the road.

The systems described above are relatively complicated and are not directly adapted for use in terrain. The systems are usually also adapted to detect objects in general, and not to distinguish specific objects.

It is thus an object of the invention to provide an improved system for preventing an autonomous vehicle from being punctured in an off-road environment.

Summary of the invention According to one aspect, the above-mentioned object is achieved by a control system for controlling a wheeled autonomous vehicle in the detection of objects according to the first independent claim. The wing system comprises one or more sensing units, each of which is adapted to sense the oncoming road path of one of the wheels of the vehicle. By identifying particularly dangerous objects such as so-called "cat skulls" in this upcoming roadway, the vehicle can steer to the side and avoid these objects. A "cat skull" in a mining environment is a fist-sized stone with sharp edges that is known for puncturing the vehicle's wheels. By using the control system to identify such objects, the vehicle can avoid having its wheels punctured and thus avoid interruptions in production because the vehicle needs to stand still and block the production line, or deviate from its work to have its wheels repaired.

The sensing units can be set according to the upcoming road surface of the vehicle's wheels. Since the objects are so small, it is sufficient to detect a narrow aisle in front of each wheel, ie a narrow width of the vehicle's roadway that corresponds to the width of the wheels. In this way, you can have simpler and cheaper sensing units that handle a dangerous mining environment in a good way.

According to another aspect, the object is achieved by a method for regulating a wheeled autonomous vehicle in the detection of objects according to the second independent claim. Preferred embodiments are defined by the dependent claims.

Brief description of the figures Figure 1 schematically shows a part of a traffic system with three autonomous vehicles shown here.

Figure 2 shows an autonomous vehicle equipped with two sensing units.

Figure 3 shows a block diagram of a control system according to an embodiment of the invention.

Figure 4 shows a flow chart of the method according to an embodiment of the invention.

Detailed Description of Preferred Embodiments of the Invention Figure 1 schematically shows a traffic system comprising three autonomous vehicles 2 traveling along a road. The arrows in the autonomous vehicles 2 show their respective direction of travel. The autonomous vehicles 2 can communicate with a control center 1 via, for example, V21 communication (Vehicle-to-Infrastructure) 3 and / or with each other via, for example, V2V communication (Vehicle-to-Vehicle) 4. This communication is wireless and can take place, for example. via a WLAN protocol (Wireless Local Area Network) IEEE 802.11, for example IEEE 802.1 1 p. However, other wireless communication methods are also possible. The command center 1 organizes the autonomous vehicles 2 and gives them assignments to perform. When an autonomous vehicle has been given an assignment, the vehicle can independently ensure that the assignment is performed. An assignment can, for example, consist of an instruction to pick up goods at a goods collection point A. The vehicle 2 then has the capacity to determine its current position, determine a road from the current position to the goods collection point A, and get there. Along the way, the vehicle must also have the capacity to give way to obstacles and handle other autonomous vehicles 2 that may have a more important task and must be given priority. In an off-road environment such as a mine, there are objects that are more dangerous to a vehicle than other objects. An example is so-called cat skulls, which are fist-sized stones with sharp edges. These objects are known for puncturing air-filled vehicle wheels, and it is desirable to avoid driving on them. In a driver-controlled vehicle, the driver can locate these objects and drive around them. The driver can see that it is just a cat's skull and can also calculate the road to drive around them. In the case of an autonomous vehicle 2, the task becomes much more difficult. The tile system 10 which will be described in the following is intended to solve this task.

Figure 2 shows an autonomous vehicle 2 in which parts of the control system 10 are shown.

The vehicle 2 has four air-borne wheels 7 and is provided with at least one sensing unit 5, 6, here two sensing units 5, 6 which are each adapted to sense the coming roadway 8 for one of the vehicle's two front wheels 7. If the vehicle 2 had been, for example, a motorcycle, a sensing unit 5 could have been sufficient. Each sensing unit 5, 6 is adapted to generate a sensing signal s1, sz which indicates object 9 in the coming roadway 8. An object 9 in the form of a cat skull is illustrated here in the roadway 8 which will to be detected by the sensing unit 5.

The forthcoming roadway 8 for the wheels 7 of the vehicle is the space the wheels 7 of the vehicle 2 will occupy along the coming road taking into account the width w of the wheels 7 and how the vehicle 2 will drive. Preferably, the roadways 8 for the front wheel pair are sensed, but it is also possible to sense the roadway for a rear wheel 7 or the roadways for two rear wheels. A wheel 7 may also comprise two or more wheels sitting together. The width of the wheels 7 is as shown in figure w, and this width needs to be sensed a bit further ahead in the travel path 8 in order to be able to drive around an object 9 in the path of the vehicle. According to one embodiment, the width w is between 20-100 cm. The sensing units 5, 6 are preferably placed high up on the vehicle 2 in order to have as large an angle to the roadway 8 as possible in order to be able to sense the roadway 8 with greater precision. The sensing units 5, 6 are preferably set to sense the roadway 8 a length I 5-30 m in front of the vehicle 2, according to an embodiment 10 m in front of the vehicle 2. The sensing units 5, 6 generate information from which at least one distance d1, k1 to the object 9 can determined. Figure 3 shows a block diagram of the control system 10. In addition to at least one sensing unit 5, 6 as described with reference to Fig. 2, the reflecting system 10 also comprises an analysis unit 11 which is adapted to receive the sensing signal or sensing signals sj, sz from the sensing units 5, 6. The analysis unit 11 is further adapted to determine at least one distance dj, kj to the object 9, compare the distance dj, kj with a reference distance djejdjdjejkj and determine at least one object profile based on the result of the comparison. An object profile for an object 9 in each roadway 8 can thus be determined. This object profile can then be used to distinguish a cat's skull from, for example, a hole in the ground. The difference distances djejdhdjejkj can for example be saved on a memory 14 connected to the analysis unit 11.

According to one embodiment, the sensing signal sj, sz also indicates distance dj, kj to the roadway 8 for the vehicle wheel 7 when no object 9 has been sensed. The analysis unit 11 can then be adapted to determine a reference distance djejjjj, djejkj which is an average value of a plurality of previously determined distances dj, kj to the road surface of the vehicle 7 wheels 7 when no object 9 is sensed. In this way, reference distances djejjjj, djejkj to the roadway 8 can be determined which the analysis unit 11 can use to compare with determined distances dj and kj, and thus obtain a difference between a reference distance djejjjj, djejkj and the determined distances dj, kj. If there is a difference, this difference can be used to create an object profile.

The analysis unit 11 is further adapted to compare the object profile with at least one reference object profile, and to determine a result signal sjj, sj2 which indicates the result of the comparison. A reference object profile is a profile that identifies a particular object. According to one embodiment, the object reference profile characterizes a fist-sized stone, for example a cat skull. The Fleferens object profile includes predetermined characteristics of the particular object, for example a height within a certain range and / or width within a certain range. If there is a match between the object profile and the reference object profile, it is indicated in the result signals sjj, sjg for each object. The mirror system 10 also comprises a control unit 12 which is adapted to receive said result signal sn, sfg and determine one or more control signals scom, according to predetermined rules based on the result of the comparison or comparisons. According to one embodiment, the predetermined rules include rules for how the vehicle 2 is to be regulated so that the vehicle's wheel 7 avoids a defined object 9 in the roadway 8. The predetermined rules include vehicle-specific determinations such as the distance between wheel pairs etc. To determine how the vehicle 2 is to be regulated in order to avoid an object 9, the control unit 12 can be adapted to receive a path signal sb which indicates the coming path of the vehicle 2.

The travel signal sb may for example comprise a trajectory for the vehicle 2. Through information about the trajectory for the vehicle 2, the distance to the object 9 and for example the distance between the front wheels 7 and the width w of the wheels 7, the control unit 12 can determine how the vehicle must be regulated to avoid the object 9 If an object 9 has been identified as a cat's skull, the vehicle 2 must drive around the cat's skull. One or more control signals soon "are then determined to move the vehicle 2 around the object 9.

The control unit 12 is further adapted to send control signals scom, to a control system 13 in the vehicle 2, the vehicle 2 being controlled accordingly. The control system 13 can for instance be a control system 13 which regulates the steering of the vehicle 2. When the control system 10 comprises two or more sensing units 5, 6 which are adapted to sense the coming roadway 8 for each of the wheels 7 of the vehicle 2, the control unit 12 is adapted to determine control signals which are also based on objects 9 in the coming roadway 8 for the other of the wheel 7 of the vehicle 7. By the second of the wheel 7 of the vehicle 2 is meant here the second wheel of a front pair of wheels.

The sensing units 5, 6 as partially described before, may comprise for example one or more lasers each which can determine distance d1, k1 to the roadway 8 or an object 9. The laser may be a scanning laser which is adapted to scan the width w of the roadway 8. From this scan, a plurality of substantially simultaneous distances d1-dx, k1-kX to object 9 can be determined. The sensing units 5, 6 may instead comprise a plurality of lasers each, which are placed in a row or in a matrix. For example, three lasers can be placed in width and set so that together they sense the width w a certain length I in front of the vehicle 2. According to an embodiment, the analysis unit 11 is then adapted to determine a plurality of substantially simultaneous distances d1-dx. , k1-kx to the object 9, and comparing the simultaneous distances d1-dx, k1-kX with one or more reference values drefdfdrefdx, drem-drefkx and determining an object profile based on the results of the comparisons.

A time series of fixed distances can also be used. The analysis unit 11 is then adapted to determine an object profile based on a time series of results of comparisons from a time series of distance determinations. The reference object profile then has equivalent determinations for what the reference object looks like, in order to be able to be compared with the determined object profile. If the lasers are placed in a matrix, for example 2x2 or 3x3, they can directly sense an object 9 in three dimensions. Then an object profile in 3D can be determined, which can be compared with a reference object profile in 3D.

The sensing units 5, 6 may instead comprise cameras from whose images one can identify objects 9 and determine the distance d1, k1 to these. In order to be able to determine depth, however, two cameras are needed in each sensing unit 5, 6.

As previously described, the control unit 12 may be adapted to receive a roadway signal sb indicating the vehicle's upcoming roadway 8. The roadway signal may comprise a trajectory for the vehicle 2 which can be determined taking into account how the vehicle 2 is steered, for example steering wheel adjustment or wheel adjustment. The trajectory can, for example, indicate where the center line of the vehicle 2 in the direction of travel should be, and the control unit can then be adapted to determine the trajectories 8 of the wheels. the control unit 12 must then be adapted to independently determine trajectories 8 for the wheels 7 of the vehicle 2.

The control unit 12 is further adapted to determine at least one setting for at least one sensing unit 5, 6 based on the upcoming roadway 8 of the vehicle 2, so that the sensing unit 5, 6 senses the coming roadway 8 for one of the vehicle wheels 7. To determine one or more settings for the sensing units 5, 6, the control unit 12 can use the determined trajectories for the vehicle wheels 7, the selected distance 1 in front of the vehicle 2, and at what height of the vehicle 2 the sensing devices 5, 6 are located. By geometric calculations, the control unit 12 can then determine settings for the sensing units 5, 6 so that they sense in a region in the coming travel path 8 of the wheels 7 a certain distance I in front of the vehicle 2. The respective sensing unit 5, 6 can thus be set to sense a particular path 8. The control unit 12 is further adapted to generate a setting signal sse fl, ssetg which indicates the setting and send this to the sensing unit 5, 6, the sensing unit 5 being set after the setting. Each sensing unit 5, 6 is thus adapted to be able to be set so that it senses a dedicated roadway 8.

In this way, the control system 10 can constantly sense the roadway 8 where the wheels 7 of the vehicle will be placed.

Depending on how the vehicle 2 turns or plans to turn, the sensing units 5, 6 will thus be set accordingly. This setting can be visible to other road users and in this way they are informed about where the vehicle 2 is heading.

The analysis unit 11 and the control unit 12 can for instance be comprised of a computer in the vehicle 2, or in a control unit (ECU - Electronic Control Unit). The wing system 10 preferably includes a processor and memory 14 for performing the methods described herein. The wing system 10 is adapted to communicate with different units and systems in the vehicle 2 via one or more different networks in the vehicle 2, such as a wireless network, via CAN (Controller Area Network), LIN (Local Interconnect Network) or Flexray etc.

The invention also relates to a method for controlling a wheeled autonomous vehicle 2 in the detection of object 9, which will now be described with reference to the flow chart in Figure 4. The method comprises a first step A1) to detect the coming roadway 8 for at least one of the wheels 7 of the vehicle 2. This can be done with one or more sensing units 5, 6 as previously described in connection with the control system 10. In a second step A2) at least one distance dj, k1 to an object 9 in the roadway 8 is determined. In a third step A3), the distance d1, k1 is compared with a reference distance drerr, drerrr, and an object profile is determined based on the result of the comparison. In a fourth step A4), the object profile is compared with a reference object profile, and a result of the comparison is determined.

The reference object profile can, for example, characterize a fist-sized stone, a so-called cat skull. If the object profile includes one or more features of a cat skull, the result of the comparison is that the object profile matches the reference object profile and a cat skull has been identified. In order to avoid the identified cat skull, the method in a fifth step A5) comprises determining one or more control signals seerrrr according to predetermined rules based on the result of the comparison. The rules include rules for how the vehicle 2 is to be regulated to avoid the defined object 9, which was previously explained in connection with the control system 10. In a sixth step A6) they are sent one or more control signals seerrrr to a control system 13 in the vehicle 2, whereby the vehicle 2 controlled accordingly.

According to one embodiment, the method of determining a reference distance comprises drerrr, drerrr which is an average value of a plurality of previously determined distances dr, kr to the roadway 8 of the vehicle wheel 7 when no object 9 is sensed. In this way one can then obtain difference values between a reference distance drerer, drerrr and determined distances to an object 9.

According to one embodiment, the method in step A2) comprises determining a plurality of simultaneous distances dr-dx, kr-kx to the object 9, and in step A3) comparing said plurality of simultaneous distances dr-dx, kr-kx with one or more reference values -drerex, drerrrr-drerrx and determine an object profile based on the results of the comparisons.

In this way, a more secure analysis can be made. According to a further embodiment, step A3 comprises determining an object profile based on a time series of results of comparisons derived from a time series distance determinations. In this way, a 2D or 3D object profile can be built up, and compared with an equivalent reference object profile. In this way, an even more reliable analysis can be made.

According to one embodiment, the method comprises the steps of: 11) receiving information about the upcoming travel path of the vehicle 2; B2) determining at least one setting for at least one sensing unit 5, 6 based on the coming path of the vehicle 2, so that each one or more sensing units 5, 6 senses the coming path 8 for one of the wheels 7 of the vehicle; B3) send the setting to the sensing unit or sensing units 5, 6, the sensing unit or sensing units 5, 6 being set after the setting. In this way, the sensing units 5, 6 can always sense the travel path 8 for the wheels 7 of the vehicle 2.

The invention also relates to a computer program P in an autonomous vehicle 2, wherein the computer program P comprises program code for causing the control system 10 to perform the steps according to the method. Figure 3 shows the computer program P as part of the memory 14. The computer program P is thus stored on the memory 14. The memory 14 is connected to the analysis unit 11, and when the computer program P is executed by the analysis unit 11, at least parts of the methods which have described herein. The invention further comprises a computer program product comprising a program code stored on a computer readable medium for performing the method steps described herein, when the program code is executed on the control system 10. The memory 14 may also be connected to the control unit 12.

The present invention is not limited to the above-described preferred embodiments. Various alternatives, modifications and equivalents can be used.

The above embodiments are therefore not to be construed as limiting the scope of the invention as defined by the appended claims. 12

Claims (18)

10 15 20 25 30 536 983 Patent claims Control system (10) for controlling a wheeled autonomous vehicle (2) when detecting objects (9) and wherein the autonomous vehicle is intended to operate in an off-road environment, the control system (10) comprising - at least one sensing unit (5, 6 ) each adapted to sense the oncoming roadway (8) for one of the wheels of the vehicle (7), and to generate a sensing signal S1, sz indicating objects (9) in the oncoming roadway (8); an analysis unit (11) adapted to receive said sensing signal s1, sz and determine at least one distance d1, k1 to the object (9), compare the distance d1, k1 with a reference distance d, efd1, d, efk1 and determine at least one object profile based on the result of the comparison; the analysis unit (11) is further adapted to compare the object profile with a reference object profile, and determine a result signal sn, sfz which indicates the result of the comparison; a control unit (12) adapted to receive said result signal s, and determining one or more control signals which, according to predetermined rules based on the result of the comparison, the control unit (12) is further adapted to send said control signals scomr to a control system (13 ) in the vehicle (2), the vehicle (2) being steered accordingly. Control system (10) according to claim 1, wherein the sensing signal s1, sz also indicates distance d1, k1 to the road surface (8) of the vehicle wheels (7) when no object (9) is sensed; wherein the analysis unit (11) is adapted to determine a reference distance dæfdh drew which is an average value of a plurality of previously determined distances d1, k1 to the road surface of the vehicle wheels when no object is sensed. A control system (10) according to claim 1 or 2, wherein the analysis unit (11) is adapted to determine a plurality of simultaneous distances d1-dx, k1-kx to the object (9), the analysis unit (11) being adapted to compare said plurality of simultaneous distance d1-dx, k1-kx with one or more reference values dæfdfdfefdx, dfe fl q-dfefkx and determine an object profile based on the results of the comparisons. 13 10 15 20 25 30 536 983 A control system (10) according to any one of the preceding claims, wherein said analysis unit (11) is adapted to determine an object profile based on a time series of results of comparisons from a time series of distance determinations. Control system (10) according to any one of the preceding claims, wherein the control unit is adapted to receive a travel path signal sb indicating the coming road path of the vehicle, the control unit (12) being adapted to determine at least one setting for said at least one sensing unit (5, 6). based on the oncoming road path of the vehicle, so that the sensing unit (5, 6) senses the oncoming road path (8) for one of the wheels (7) of the vehicle, the control unit (12) being adapted to generate a setting signal. sm; indicating said setting and sending it to the sensing unit (5, 6), the sensing unit (5, 6) being set after the setting. A control system (10) according to any one of the preceding claims, wherein said at least one sensing unit (5, 6) comprises at least one laser. A control system (10) according to any one of the preceding claims, wherein said object reference profile characterizes a fist-sized stone. Control system (10) according to any one of the preceding claims, wherein said predetermined rules comprise rules for how the vehicle (2) is to be regulated in order for the vehicle wheels (7) to avoid a defined object (9) in the roadway (8) of the vehicle wheels ( 7). Control system (10) according to any one of the preceding claims, comprising two sensing units (5, 6) adapted to sense the oncoming road path for each one of the wheels of the vehicle, the control unit (12) being adapted to determine control signals and, also based on objects (9) in the oncoming lane (8) for the other of the vehicle's wheels (7). 14 10 15 20 25 30 536 983 Method for controlling a wheeled autonomous vehicle (2) when detecting objects (9) and wherein the autonomous vehicle is intended to operate in an off-road environment, which comprises the steps of: - sensing the oncoming roadway (8) for at least one of vehicle wheels (7); - determining at least one distance d1, k1 to an object (9) in the roadway (8); - comparing the distance d1, k1 with a reference distance drew, drew, and determining an object profile based on the result of the comparison; - comparing the object profile with a reference object profile, and determining a result of the comparison; - determining one or more control signals scom, according to predetermined rules based on the result of the comparison; - sending said one or fl your control signals smm, to a control system (13) in the vehicle (2), the vehicle (2) being controlled accordingly. A method according to claim 10, comprising determining a reference distance drew, dæ fl d which is an average of a plurality of previously determined distances d1, k1 to the road surface (8) of the vehicle wheels (7) when no object (9) is sensed. A method according to claim 10 or 11, comprising determining a plurality of simultaneous distances d1-dx, k1-kx to the object (9), comparing said plurality of simultaneous distances d1-dx, k1-kx with one or more reference values dfefm-drefdx , dfem-dfefkx and determine an object profile based on the results of the comparisons. A method according to any one of claims 10 to 12, comprising determining an object profile based on a time series of results of comparisons derived from a time series distance determinations. A method according to any one of claims 10 to 13, comprising the steps of - receiving information about the coming path of the vehicle (2); - determining at least one setting for at least one sensing unit (5, 6) based on the oncoming road path of the vehicle, so that the sensing unit (5, 6) senses the oncoming road path (8) for one of the wheels (7) of the vehicle; 15 10 15 536 983 - send the setting to the sensing unit (5, 6), the sensing unit (5, 6) being set after the setting. A method according to any one of claims 10 to 14, wherein said object reference profile characterizes a fist-sized stone. A method according to any one of claims 10 to 15, wherein said predetermined rules comprise rules for how the vehicle (2) is to be regulated so that the wheels (7) of the vehicle (2) avoid a defined object (9) in the roadway (8) of the vehicle. wheels (7). A computer program (P) in an autonomous vehicle, said computer program (P) comprising program code for causing a control system (10) to perform the steps of any of claims 10 to 16. A computer program product comprising a program code stored on a computer readable medium for performing the method steps according to any one of claims 10 to 16, when said program code is executed on a control system (10). 16