Collision avoidance procedures

Abstract

Method for avoiding a collision between an agricultural vehicle and a potential collision object on agricultural land, comprising the steps of: - detecting (S1a, S1b) objects on the agricultural land using a laser scanner and a radar device, wherein the detection step (S1a, S1b) includes detecting (S1a) vegetation on the agricultural land using the laser scanner, - classifying (S2) the agricultural land into collision zones and traversable zones for the agricultural vehicle based on the objects detected by the laser scanner and the radar device, wherein the classification step (S2) is performed depending on a current work operation being carried out by the agricultural vehicle, and - intervening (S3) in the driving dynamics of the agricultural vehicle when a predetermined,future movement path of the agricultural vehicle passes through one of the classified collision zones.

Description

The invention relates to a method for avoiding a collision of an agricultural utility vehicle with a potential collision object on an agricultural working surface.The invention also relates to a control device which is configured to carry out such a method, and to a collision avoidance system having such a control device.It is known to navigate an agricultural utility vehicle on an agricultural working surface by means of satellite-assisted positioning methods. Differential GPS is used for this purpose, for example, in order to increase the precision of the navigation. In addition, it is also known to detect obstacles using environment detection sensors arranged on an agricultural utility vehicle. DE 10 2016 118 227 A1 describes an image analysis system for agricultural working machines.In one aspect, the invention relates to a method for avoiding a collision of an agricultural utility vehicle with a potential collision object on an agricultural working surface.A collision may include a collision of the agricultural utility vehicle with the potential collision object. The agricultural utility vehicle may be in motion when performing the method. The potential collision object may be a stationary object or a moving object.The agricultural utility vehicle may be a vehicle for managing an agricultural working surface. The agricultural utility vehicle can be designed as an autonomous utility vehicle or as an automatically operable utility vehicle. For example, the agricultural utility vehicle can be a tractor or an agricultural sprayer. The sprayer may also be an agricultural sprayer or an agricultural syringe. The agricultural working surface can be a working surface on which an agricultural utility vehicle can operate. For example, the agricultural working surface can be a field or a meadow.The potential collision object may be an object that cannot be traveled over by the agricultural utility vehicle. In other words, the potential collision object can be an object which can be damaged or cause damage to the agricultural utility vehicle in the event of a collision with the agricultural utility vehicle itself.As one step, the method comprises detecting objects on the agricultural working surface with a laser scanner and a radar device. The method can thus comprise, as a substep, capturing objects on the agricultural working surface with a laser scanner. The method can also comprise, as a further substep, detecting objects on the agricultural working surface using a radar device. Objects on the agricultural working surface can thus be detected in parallel or redundantly by a laser scanner and / or a radar device. The object detection may be synchronized. Furthermore, in the environment of the agricultural utility vehicle, an object can be detected by one of the laser scanner and the radar device and another object by the other of the laser scanner and the radar device.The objects may have a potential collision object. The objects can be objects connected to the agricultural working surface, which objects can be plants rooted in the agricultural working surface or growing on the agricultural working surface. The objects can be objects which move on the agricultural working surface or which rest on it. In other words, the objects may include dynamic objects and / or static objects.The objects can have vegetation, rock, living beings or buildings, for example. The vegetation can be crop plants, crop plants or crop material. The living beings can be animals or humans. Thus, the vegetation can be foliage, grass or harvested hay. The hay can be located, for example, as vegetation in a swath on the agricultural working surface. The vegetation can be objects which can be traveled over by the agricultural utility vehicle. By contrast, rocks, living things, trees or buildings can be a potential collision object. In one example, the potential collision object is a rock (chunk).A potential collision object can be recognized in the detected objects by means of methods of pattern recognition or template matching. The laser scanner and / or the radar device can each capture a point cloud in which the potential collision object can be detected. For this purpose, additional information from the measurement data of the laser scanner and / or of the radar device can also be used. For example, reflectivity values of received radar radiation and / or intensity values of received laser radiation can be used. Alternatively or additionally, a potential collision object in the detected objects can be detected by a comparison with a detected ground surface of the agricultural working surface or a predetermined terrain model of the agricultural working surface. Alternatively or additionally, a potential collision object in the detected objects can be detected by comparing the environment of the agricultural utility vehicle detected at an earlier point in time. For a recognition of vegetation, what is described for the recognition of potential collision objects can be applied analogously.As a further step, the method comprises classifying the agricultural working surface into collision areas and areas which can be traveled over by the agricultural utility vehicle. This step is carried out on the basis of the objects detected by the laser scanner and the radar device.The classification can comprise a division, a so-called "mapping" or a mapping of the agricultural useful surface into collision areas and drivable areas. In other words, in the step of classifying, a map of a surrounding area of the agricultural utility vehicle detected by the laser scanner and the radar device can be created, which map has collision areas and areas that can be traveled over for the agricultural utility vehicle. It is thus possible to distinguish between regions which are currently drivable and regions in which obstacles are located. The map can be generated based on raw data and / or point clouds of the laser scanner and / or of the radar device.The map may function as a travelability map. For creating the map, objects detected with the laser scanner and / or with the radar device can be combined. A first travelability map can be created based on the objects detected with the laser scanner. A second travelability map can also be created based on the objects detected with the radar device. The two maps can be combined and, based on the detected objects, the collision areas and the areas that can be traveled over by the agricultural utility vehicle can be defined or mapped in the map.Collision areas can be areas of the agricultural working surface in which at least one potential collision object has been detected or recognized by the laser scanner and / or the radar device. The areas that can be traveled over for the agricultural utility vehicle can be areas in which no potential collision object has been detected by the laser scanner and / or the radar device. In other words, in the step of classifying, the agricultural working area is divided into areas with potential collision objects and areas without potential collision objects. The passable areas may include unobstructed areas and / or areas with passable obstacles. Furthermore, regions can be derived on the basis of a grid which can be placed over the agricultural working surface. Areas can also be spanned by cells, for example by square cells, on the agricultural working surface.As a further step, the method has an intervention in the driving dynamics of the agricultural utility vehicle. The intervention is carried out when a predetermined future movement path of the agricultural utility vehicle passes through one of the classified collision regions. In addition, the intervention can also be carried out when the predetermined future movement path of the agricultural utility vehicle runs at a distance from one of the classified collision areas which is less than a predefined distance.The intervention can comprise an intervention in the transverse dynamics and / or in the longitudinal dynamics of the agricultural utility vehicle. The intervention in the transverse dynamics can comprise controlling or steering the agricultural utility vehicle. The intervention in the longitudinal dynamics can comprise a braking or an acceleration of the agricultural utility vehicle.The predetermined future movement path of the agricultural utility vehicle can have at least a section of a predefined route or trajectory of the agricultural utility vehicle to be traveled on the agricultural useful surface. Such a movement path can be precomputed on the basis of a positioning or localization of the agricultural utility vehicle on a map. Positioning or locating of the agricultural utility vehicle can be carried out with a satellite-assisted positioning system, for example a GNSS system. A GNSS receiver can be arranged for this purpose on the agricultural utility vehicle. The movement path can serve for managing the agricultural working surfaceAlternatively or additionally, the predetermined future movement path of the agricultural utility vehicle can be derived from a sensor-based detection of objects on the agricultural useful surface. Sensor-based recognition of objects can be carried out with a sensor-based tracking system. The sensor-based track system can be arranged on the agricultural utility vehicle. For example, the agricultural utility vehicle can thus follow a vegetation edge, for example a section line on a field, in the future.Within the scope of the invention, an agricultural working surface can thus be divided into drivable and non-drivable regions with a multisensor system which has at least the laser scanner and the radar device. The multisensor system can be arranged on the agricultural utility vehicle.Based on this, a travelability map in the environment of the agricultural utility vehicle can be generated with the invention, in which map the agricultural utility vehicle can be navigated independently of a predetermined movement path. It is also possible to generate a driveability map in the environment of the agricultural utility vehicle, in which map the utility vehicle can drive over detected objects. The method can therefore also be carried out as a method for navigating the agricultural utility vehicle. In this case, the agricultural utility vehicle can be controlled into drivable regions with objects that can be driven over.Within the scope of the invention, the agricultural utility vehicle can thus automatically react to a potential collision object along a predetermined movement path and drive around the latter. The agricultural utility vehicle can deviate from the predetermined movement path and return to the predetermined movement path again after passing around the potential collision object.Thus, with the invention, autonomous operation of an agricultural utility vehicle or so-called "precision farm" can be advantageously improved. The method can be used to react automatically to non-crossover obstacles by a control intervention in the vehicle drive of the agricultural utility vehicle. A further advantageous effect of the invention can be seen in the fact that operation of an agricultural utility vehicle on an agricultural working surface can be more reliable and efficient. If a potential collision object is a living being, the object can also be located more safely on a surface managed by an agricultural utility vehicle using the invention. The agricultural utility vehicle may further be remotely controlled for performing the method.The step of detecting comprises detecting vegetation on the agricultural working surface with the laser scanner. Alternatively or additionally, the step of detecting can comprise detecting a floor surface of the agricultural working surface with the laser scanner. The step of detecting may also comprise detecting potential collision objects which are not surrounded by vegetation. Objects not shaded by vegetation can therefore also be detected in the environment of the agricultural utility vehicle. The laser scanner can be arranged on the agricultural utility vehicle.According to a further embodiment of the method, the step of detecting comprises detecting objects which are at least partially surrounded by vegetation with the radar device. The radar device can be designed as an ultra-wide band radar device (UWB radar). Such a radar device can be designed to emit radar radiation which can penetrate vegetation on the agricultural working surface. With such a radar device, objects can thus be detected which are at least partially surrounded by vegetation. In other words, the radar device can detect objects shaded from vegetation in a detection direction of the radar device. Such objects can be potential collision objects which are present in vegetation on the agricultural working surface.Thus, potential collision objects which are not visible to a driver of the agricultural utility vehicle or are hidden by vegetation and cannot be detected by the laser scanner can be detected in an advantageous manner by means of the radar device. For example, potential collision objects in a swath can thus be detected before the swath is recorded. In a further example, potential collision objects in a meadow, a cereal field or a maize field can thus be detected before they are mowed or harvested. Damage to the agricultural machine, damage to agricultural equipment, damage to persons or wild damage can thus be avoided. Alternatively or additionally to the laser scanner, the radar device can be arranged on the agricultural utility vehicle.According to a further embodiment of the method, the step of classifying comprises assigning non-binary driveability values to the potential collision regions. The non-binary driveability values may sometimes have two binary driveability values (one for driveability and zero for non-driveability). The non-binary override values may further include normalized values. The step of classifying can thus also comprise quantifying the difficulty for the agricultural utility vehicle in relation to the driveability of a region of the agricultural useful surface.Assigned driveability values can indicate the probability of a collision with a potential collision object imminent when driving over a collision area. The assigned driveability values may have probability values, which may indicate the probability of a collision imminent. In other words, the assigned driveability values can indicate a driveability level of a potential collision area.According to a further embodiment of the method, the driveability values are scaled as a function of a detection time by the laser scanner and / or by the radar device. In other words, more reliable or higher driveability values can be generated if a region has been detected longer or more frequently. Thus, the driveability values can be generated depending on how long a region has been in a detection range of the laser scanner and / or of the radar device.According to a further embodiment of the method, the step of classifying is carried out as a function of at least one detected dimension of the potential collision object. The dimension or a volume of the potential collision object can be detected by the laser scanner and / or the radar device. Thus, areas of the agricultural working surface can be classified as collision areas if a detected dimension of a potential collision object in these areas exceeds a predefined maximum dimension.According to a further embodiment of the method, the step of classifying is carried out as a function of at least one dimension of the agricultural utility vehicle. The dimension can also relate to the agricultural utility vehicle as such or to a component of the agricultural utility vehicle. The dimension can have, for example, a width, a length and / or a height of the agricultural utility vehicle. In another example, the dimension may include a wheel diameter of a front and / or a rear wheel of the agricultural utility vehicle. Depending on at least one dimension of this type, for example one of the wheel diameters, the assigned driveability values can be defined.According to a further embodiment of the method, the step of classifying is carried out as a function of at least one dimension of an agricultural device arranged on the agricultural utility vehicle. The agricultural device can be, for example, an add-on sprayer or a baler. The dimension may include, for example, a width, a height and / or a length of the agricultural implement. Thus, the assigned driveability values can be defined as a function of a dimension of the agricultural utility vehicle and / or a dimension of an agricultural implement arranged on the same.The classifying step is carried out as a function of a work process currently carried out by the agricultural utility vehicle. The operation may be any operation suitable for managing an agricultural crop. For example, the work operation can comprise mowing, turning, windrowing or harvesting. Alternatively or additionally, the step of intervening in the driving dynamics of the agricultural utility vehicle can also be carried out as a function of the work process currently carried out by the agricultural utility vehicle.According to a further embodiment of the method, the step of intervention comprises the initiation of emergency braking of the agricultural utility vehicle. Emergency braking can be initiated automatically by the agricultural utility vehicle. Thus, in one example, a reaction can be made to a living being residing in a corn field by stopping in good time before the latter, although the living being is not perceptible to the driver and cannot be detected by the laser scanner.According to a further embodiment of the method, the step of intervention comprises a deviation from the predetermined future movement path of the agricultural utility vehicle. The deviating may comprise correcting the movement path. The method can therefore be used to react to a detected potential collision object by a driving dynamics intervention independently of the predetermined future movement path. The step of engaging may comprise controlling the agricultural utility vehicle along a correction path which deviates from the predetermined future movement path. The correction path can initially lead away from the movement path, then run parallel to the latter and open again into the latter.In a further aspect, the invention relates to a control device which is configured to carry out method steps according to the preceding aspect.The control device can have interfaces for reading in detection data of the laser scanner and of the radar device. The detection data can comprise information about objects of the agricultural working surface. The information can comprise position information relating to the relative position of the objects with respect to the agricultural utility vehicle. The control device can also have a classifier for classifying the agricultural working surface according to the step of classifying the preceding aspect. The control device can furthermore have an interface for outputting an intervention command for intervening in the driving dynamics of the agricultural utility vehicle according to the step of intervening of the preceding aspect. The intervention command can be output when a predetermined future movement path of the agricultural utility vehicle passes through a collision region classified with the classifier.In a further aspect, the invention relates to a collision avoidance system having a laser scanner, a radar device and a control device according to the preceding aspect. The collision avoidance system may function as a system parallel to a navigation system of an agricultural utility vehicle. In other words, the collision avoidance system may be a system that can operate substantially independently of a navigation system. FIG. 1 shows a flow diagram of method steps of a method for avoiding a collision of an agricultural utility vehicle with a potential collision object on an agricultural working surface according to one embodiment of the invention.FIG. 1 shows method steps S 1, S 2, S 3 in a temporal sequence. Method steps S 1, S 2, S 3 are carried out continuously in order to avoid a collision of an agricultural utility vehicle with a potential collision object on an agricultural useful surface.The first step S 1 is divided into substeps S 1 aand S 1 b. In the first substep S 1 a, objects on the agricultural working surface are detected by means of a laser scanner. In the second substep S 1 b, the same or different objects on the agricultural working surface are detected with a radar device. In the first step S 1, the objects detected with the laser scanner and the radar device are additionally fused or combined.In a further step S 2, the agricultural useful area is classified into collision areas and into areas which can be traveled over for the agricultural utility vehicle, based on the objects detected in step S 1. At least one object detected in step S 1 is located in a collision area, which represents a potential collision object.In a further step S 3, a driving dynamics intervention takes place. In this step, the driving dynamics of the agricultural utility vehicle are taken into account if a predetermined future movement path of the agricultural utility vehicle passes through a collision region classified in step S 2. A collision with a potential collision object residing therein can thus be avoided.Reference numerals denote reference numeralsS 1 Object detection S 1 a Object detection Laser scanner S 1 b Object detection Radar device S 2 Classification S 3 Driving dynamics intervention

Claims

Method for avoiding a collision of an agricultural utility vehicle with a potential collision object on an agricultural useful surface, having the steps: - detecting (S1a, S1b) objects on the agricultural useful surface using a laser scanner and a radar device, wherein the step of detecting (S1a, S1b) comprises detecting (S1a) vegetation on the agricultural useful surface using the laser scanner, - classifying (S2) the agricultural useful surface into collision regions and regions which can be traveled over for the agricultural utility vehicle on the basis of the objects detected using the laser scanner and the radar device, wherein the step of classifying (S2) is carried out as a function of a working process currently carried out by the agricultural utility vehicle, and - Intervention (S3) in the driving dynamics of the agricultural utility vehicle if a predetermined future movement path of the agricultural utility vehicle runs through one of the classified collision areas.Method according to claim 1, wherein the step of detecting (S1a, S1b) comprises detecting (S1b) objects at least partially surrounded by vegetation with the radar device.The method of any preceding claim, wherein the step of classifying comprises associating non-binary passability values with the potential collision areas.Method according to one of the preceding claims, in which the classification step (S2) is carried out as a function of at least one dimension of the agricultural utility vehicle.Method according to one of the preceding claims, in which the classification step (S2) is carried out as a function of at least one dimension of an agricultural appliance arranged on the agricultural utility vehicle.Method according to one of the preceding claims, in which the step of intervention (S3) comprises the initiation of emergency braking of the agricultural utility vehicle.Method according to one of the preceding claims, in which the step of intervention (S3) comprises a deviation from the predetermined future movement path of the agricultural utility vehicle.A control device comprising means for performing the steps of a method according to any of the preceding claims.A collision avoidance system comprising a laser scanner, a radar apparatus and a controller according to claim 8.

Images