Determining a relative movement

Abstract

Device (16) for determining a relative movement between an environmental sensor (12) and an object (14), comprising: an input interface (20) for receiving a target point list of the environmental sensor with information on Doppler velocities and positions of target points on the object; a translation unit (22) for determining a translation speed of the object based on the target point list; a rotation unit (24) for determining a rotation of the object based on the target point list and the translation speed; and an output unit (26) for outputting the translation speed and the rotation of the object, wherein the translation unit is configured to determine the translation speed based on translation modeling over a plurality of time steps and a state estimation and / or the rotation unit is configured to determine the rotation based on rotation modeling over a plurality of time steps and a state estimation.

Description

The present invention relates to a device for determining a relative movement between an environment sensor and an object. The present invention further relates to a system and a method for determining a relative movement between an environment sensor and an object.Modern vehicles (cars, transporters, trucks, motor cycles etc.) have a multiplicity of sensors which make information available to the driver and control individual functions of the vehicle in a partially or fully automated manner. Sensors are used to detect the surroundings of the vehicle and other road users. Based on the captured data, a model of the vehicle environment may be generated and responsive to changes in that vehicle environment.An important sensor principle is radar technology. Most radar sensors used today in the vehicle sector operate as multi-pulse radar sensors (also referred to as chirp sequence radar sensors), in which a plurality of frequency-modulated pulses are emitted at short intervals. The radar sensors typically comprise a plurality of transmitting and receiving elements (antenna array) which form virtual channels (Rx / Tx antenna pairs). After preprocessing, the radar sensor then periodically provides a radar target list (also referred to as a point cloud) for further processing. This radar target list comprises, for the detected targets, in particular the parameters distance, radial or Doppler velocity and (if available) azimuth and elevation angles and forms the basis for environment recognition.Another relevant sensor principle is lidar technology (light detection and ranging). A lidar sensor is based on the emission of light signals and the detection of the reflected light. A distance from the location of the reflection can be calculated by means of a propagation time measurement. In addition, in modern lidar sensors, it is also possible to ascertain a relative or Doppler velocity (so-called FMCW lidar). A corresponding lidar target list may be provided in a manner comparable to a radar sensor.In the field of driver assistance systems and (partially) autonomous driving, the determination of a relative position of another object with respect to the own vehicle or to the position of the environment sensor represents an important problem. On the one hand, the host vehicle can be localized with respect to the environment. On the other hand, other objects may be located in the environment, such as traffic signs or other vehicles.In order to enable evaluation of the sensor data over a plurality of time steps, it is often necessary in this case to establish a relationship between a target list (target point list) which was determined at a first point in time and a target list which was determined at a second point in time (scan matching). This reference forms the basis for a tracking of the object or of its own position. In order to establish this reference, in particular an estimation or determination of a relative movement between the environment sensor and the object can take place and a transformation can be determined. The transformation usually comprises a rotation matrix and a translation vector. A widely used approach is the Iterative Closest Point (ICP) approach, which provides an iterative solution for finding point correspondences between different scans and determines an optimum transformation on the basis of the found correspondences. Here, the positions of the different points are considered.US 2018 / 0356517 A1 discloses a method for determining a yaw rate of a target vehicle in a horizontal plane by a host vehicle equipped with a radar system. A radar system includes a radar sensor unit configured to receive signals of the host vehicle reflected at the target vehicle. The method comprises steps of emitting a radar signal at a time instance and determining range, azimuth and rate of change for each point detection. Further, the method comprises steps of determining the values of the longitudinal and lateral components of the range rate equation of the target and the longitudinal and lateral speeds of the sensor unit or the host vehicle.A disadvantage of previous approaches for determining a relative movement is that inaccuracies can arise due to the high deviations or the sensor inaccuracies of radar or lidar sensors. Inaccuracies in scan matching can result in the downstream information processing approaches also providing erroneous results. Approaches that result in higher precisions are often significantly more computationally expensive and therefore impractical.Proceeding from this, the present invention has for its object to provide an approach for determining a relative movement between an environment sensor and an object, which approach offers high reliability and accuracy and can also be calculated efficiently. In particular, a highly accurate scan matching over a plurality of sampling times is to be made possible, in order thereby to allow self-localization and / or localization of other objects.To achieve this object, the invention relates, in a first aspect, to a device for determining a relative movement between an environment sensor and an object, having:an input interface for receiving a target point list of the environment sensor with information about Doppler speeds and positions of target points on the object;a translation unit for determining a translation speed of the object based on the target point list;a rotation unit for determining a rotation of the object based on the target point list and the translation speed; andan output unit for outputting the translation speed and the rotation of the object,wherein the translation unit is configured to determine the translation speed based on a translation modelling over a plurality of time steps and a state estimate and / or the rotation unit is configured to determine the rotation based on a rotation modelling over a plurality of time steps and a state estimate.In a further aspect, the present invention relates to a system for determining a relative movement between an environment sensor and an object, comprising:an apparatus as described above; andan environment sensor for detecting objects in an environment of a vehicle.Further aspects of the invention relate to a method embodied according to the apparatus described above and to a computer program product having program code for carrying out the steps of the method when the program code is executed on a computer, and to a storage medium on which a computer program is stored which, when it is executed on a computer, brings about execution of the method described herein.Preferred embodiments of the invention are described in the dependent claims. It is understood that the features mentioned above and those still to be explained below can be used not only in the respectively specified combination, but also in other combinations or alone, without departing from the scope of the present invention. In particular, the system, the method and the computer program product can be embodied according to the embodiments described for the device in the dependent claims.According to the invention, it is provided that a target point list of an environmental sensor is received via an input interface. In particular, data of a radar sensor or of a lidar sensor are received, which data comprise, on the one hand, a position and, on the other hand, a Doppler velocity of a plurality of sensor target points. Preferably, target point lists are continuously received at successive sampling times of the sensor. In a first step, a translation speed, i.e. a linear speed, of the object is then determined. For this purpose, in particular the Doppler speeds of the various target points on the object are evaluated. The relative speed is measured directly in this respect in the form of the Doppler speed. Only a corresponding conversion is necessary in order to calculate a linear velocity in a sensor-fixed coordinate system starting from the Doppler velocity (radial velocity) taking into account the angular position by means of a transformation. The translation speed results in a linear relative movement taking into account the sampling frequency. In a further step, a rotation of the object is then determined on the basis of the list of target points and on the basis of the translation speed. Thus, in this respect, for the determination of the rotation of the object, recourse is made to the translation speed previously determined in a separate step. The determined translation speed and rotation of the object are output and are then available for further processing, for example in a vehicle control unit.The method of the present invention can be used in particular for scan matching. In comparison with previous approaches, in which scan matching was carried out based on positions of different scan points, according to the invention a translation speed is first determined on the basis of the Doppler speed. This translation speed is then taken into account in the determination of the rotation. In this respect, the present invention is based on the idea that for a sensor which directly measures the Doppler velocity, this measured Doppler velocity can be used for the scan matching. The calculation of the linear velocity is preceded by the calculation of the rotation. As compared to previous approaches, more efficient calculation may be realized and at the same time higher accuracy may be achieved. As a result, a reliable and accurate detection of a relative movement and a relative position between the environment sensor and the object can be calculated.In a preferred embodiment, the rotation unit is designed to determine the rotation with the inclusion of the translation speed as a boundary condition. The rotation unit is preferably designed to determine the rotation based on a minimization of a cost function. It has been found that the previously determined translation speed can be taken into account as a boundary condition in further method steps. Taking into account in the form of a boundary condition is to the extent that the determined translation speed is taken into account as a specification. It is therefore attempted to find a rotation which describes the observation as well as possible if the previously determined translation speed is assumed. This results in efficient calculability.In a preferred embodiment, the rotation unit is configured to determine the rotation based on a rotation modelling over a plurality of time steps and a state estimate. Preferably, a state model of the rotation modelling does not depict an acceleration of the object. Further preferably, a particle filter is used for the state estimation. A state estimation is understood to mean, in particular, a stochastic state estimation. A system model is used to predict data of a second time step from data of a first time step. A measurement by a sensor is used to match this prediction to reality. Preferably, no acceleration of the object is mapped in the state vector of the model. This allows the length of the state vector to be shortened and the complexity of the calculation to be reduced. Preferably, a particle filter is used for the state estimation. As a result, almost all probability distributions can be mapped well. This results in a highly accurate estimate of the rotation of the object.In a preferred embodiment, the translation unit is designed to determine the translation speed based on a random sample consensus (RANSAC) algorithm. Preferably, the translation unit is configured to determine the translation speed in the form of a Doppler profile of the object over an azimuth angle. The translation speed and rotation are determined in particular in a coordinate system fixed to the sensor, starting from a position and orientation of the surrounding sensor. A Doppler profile of the object over an azimuth angle can be generated. This results in an efficient calculation possibility.In a preferred embodiment, the translation unit is designed to determine the translation speed based on translation modelling over a plurality of time steps and state estimation. A state model of the translation modelling preferably maps an acceleration of the object. Further preferably, a Kalman filter is applied for the state estimation. In the case of the translation unit, a more complex state model is used in this respect, in which the state vector comprises an acceleration of the object. In the calculation of a Kalman filter, a normal distribution is assumed, as a result of which the calculation can proceed more efficiently. This results in a faster calculation possibility. By additionally taking into account the acceleration of the object in the state vector, a higher accuracy can be achieved in the determination of the translation speed.In a preferred embodiment, the translation unit is designed to disregard outlier values in the target point list by applying a gating method. In a gating method, target points which have been produced with high probability by erroneous measurements or on the basis of special object properties are filtered out in particular. By neglecting such outlier values, the accuracy in the determination of the translation speed can be improved. A gating method is understood to mean not taking into account values outside an admissible range which results from a specific absolute or relative distance from an expected value. Improved accuracy can be realized.In a preferred embodiment, the device comprises an object recognition unit for recognizing target points belonging to an object. In this case, the translation unit is designed to determine the translation speed on the basis of the target points belonging to the object. The rotation unit is configured to determine the rotation based on the target points belonging to the object. In an upstream step, it is first determined, starting from the received target point list, which of the target points of the target point list belong to a common object. In other words, objects are detected. For this purpose, it is possible in particular to carry out a cluster analysis in which points which move uniformly are identified. By the preceding recognition of target points belonging to an object, the computing power required for determining the translation speed and the rotation can be reduced.In a preferred embodiment, the device comprises a classification unit for detecting whether the object is a static or a dynamic object, based on the target point list. The translation unit is for determining the translation speed depending on whether the object is a static or a dynamic object. A static object is in particular an object of the background. A dynamic object is in particular an object that moves relative to the background. If there is additional information about whether it is a static object or a dynamic object, this knowledge can be taken into account as additional model knowledge in the determination of the translation speed. The accuracy in determining the translation speed can be improved. This also results in a possible improvement in the accuracy and reliability when determining the rotation.In a preferred embodiment, the device comprises an accumulation unit for generating an accumulated target point list with information about positions of target points on the object based on target point lists received over a plurality of time steps and a transformation of the positions of the target points based on the determined translation speed and the determined rotation. Entries of a first target point list received at a first time are transformed based on translation and rotation. The entries of the first list of target points can thereby be transformed into the coordinate system of the second list of target points. The received destination lists do not necessarily include the same destination points at different times. New target points are also added. In order to increase the resolution, an accumulated target point list can therefore be generated, which comprises target points of a plurality of times. For this purpose, a corresponding coordinate transformation is carried out. In particular, a corresponding rotation matrix and a translation vector can be generated and used for each point. The transformation is effected in particular into the current sensor-fixed or vehicle-fixed coordinate system. The accumulation allows the resolution to be improved for further processing of the sensor data. This results in an improved possibility for evaluating environmental sensor data.In a preferred embodiment, the input interface is designed to receive the target point list from an FMCW radar sensor, which is preferably mounted on a vehicle. In particular, a frequency modulated continuous wave radar (FMCW) sensor may be used as the environment sensor. Such a radar sensor allows an accurate measurement of the angles, the distance and the Doppler velocity. The principle of FMCW radar sensors is the most widely used sensor principle in the automotive sector. The use of the sensor data of an FMCW radar sensor results in a wide employment of the device according to the invention.In a preferred embodiment, the input interface is configured to receive a first target point list of a first environment sensor with information about Doppler speeds and positions of target points on the object and a second target point list of a second environment sensor with information about positions of target points on the object. The translation unit is configured to determine the translation speed based on the first target point list. The rotation unit is configured to determine the rotation based on the second target point list. It is possible that for the determination of the translation speed, a first target point list of a first environment sensor is used, which list comprises a Doppler speed and position of the target points. For the determination of the rotation, data of a second environment sensor can be used, which only comprise the positions of target points. In this respect, two environmental sensors having different characteristics can be combined. In particular, a sensor that can measure a Doppler signal can be combined with a sensor that cannot measure a Doppler signal but can measure only a position of a target point. The data processing can be accelerated as a result. There are also possibilities for improving the accuracy by using or combining different sensor principles.In a preferred embodiment, the input interface is configured to receive the list of target points from an environment sensor mounted on a vehicle. The translation unit is designed to determine the translation speed in a two-dimensional coordinate system fixed to the sensor, which is oriented parallel to a ground plane of the vehicle. The rotation unit is designed to determine the rotation in the same coordinate system. In particular, both the translation speed and the rotation speed can be determined in a two-dimensional system. The consideration of a two-dimensional system is suitable in particular for the application in the vehicle environment. The movement of the own vehicle and of the object is limited to the movement in the ground plane. Only rotations with respect to the yaw axis of the vehicle are detected, other rotations are neglected, whereby the calculation can be substantially simplified. The environmental sensor is mounted in a vehicle. The data of the environment sensor serve as input data for autonomous or semi-autonomous control of vehicle functions or for providing additional information to a vehicle driver. The properties of the device according to the invention can be used particularly advantageously in this case, since the accuracy, reliability and calculability are improved compared to previous approaches.A relative movement is understood to mean a movement which the environment sensor and the object execute with respect to one another. In particular, on the one hand a movement of the object relative to the environment sensor and on the other hand a movement of the environment sensor relative to the object can be determined. This allows self-localization of the environment sensor or of a vehicle with environment sensor relative to its environment and localization of an object in the environment relative to the environment sensor or relative to the vehicle. An object may be a stationary or mobile object. In particular, an object can be another vehicle, a pedestrian or else an element of the environment, such as a roadway boundary or a tree trunk. An object is understood herein to mean, in particular, a rigid object. A translation speed and a rotation can be effected in particular by specifying a translation vector and a rotation matrix or by specifying a rotation speed and a linear speed. An environment sensor is in particular a sensor whose field of view comprises the environment of a vehicle. An environment sensor emits a signal and receives reflections from objects within its field of view. The field of view refers to an area within which objects can be detected. In particular, a radar sensor or a lidar sensor can be used as the environment sensor, which can measure and provide Doppler information. An environment sensor can comprise a plurality of individual sensors which, for example, enable 360° surround view and can thus record a complete image of the environment. A sensor target point is a single detection of the environment sensor. An object may have a plurality of individual sensor target points. An environmental sensor preferably outputs a sensor target list with individual detections, wherein for each individual detection, in particular the angle indications (azimuth, elevation), a distance indication (range) and a radial or Doppler velocity are indicated. A relative movement is understood to mean a movement of the object, which can be specified by a translation speed (or a translation vector) and a rotation (rotation matrix).The invention will be described and explained in more detail below with reference to some selected exemplary embodiments in conjunction with the accompanying drawings. The following are shown: FIG. 1 shows a schematic illustration of a system according to the invention; FIG. 2 shows a schematic illustration of an apparatus according to the invention; FIG. 3 shows a schematic illustration of the determination of the translation speed; FIG. 4 is a schematic illustration of the application of a gating method; FIG. 5 is a schematic illustration of the use of a particulate filter for determining rotation; FIG. 6 shows a schematic representation of an object reconstruction based on the approach of the invention; and FIG. 7 shows a schematic illustration of a method according to the invention.FIG. 1 schematically illustrates a system 10 according to the invention for determining a relative movement between an environment sensor 12 and an object 14. The system 10 comprises a device 16 for determining the relative movement and the environment sensor 12. It is understood that it is also conceivable for the device 16 and / or the environment sensor 12 to be configured separately.As shown in FIG. 1, different objects 14, such as other road users, in particular vehicles, or also static objects, such as houses or traffic signs, can be detected. In particular, the device according to the invention is suitable for use with rigid objects which are rigid in their shape. The environment sensor 12 can be, in particular, a radar or lidar sensor.The approach of the invention is based on the fact that, when determining a relative movement between the environment sensor 12 or vehicle 18 and object 14, firstly a linear relative speed (translation speed) is determined on the basis of Doppler speeds which are provided by the environment sensor 12. This translation speed then represents the basis for the determination of the rotation of the object. The relative movement of the object is completely described by the determined translation speed and rotation. The output data of the device 16 according to the invention can then be used as a basis for environment recognition and, for example, an evaluation of the environment for an autonomous driving system. In particular in the case of radar detections in which the position measurement is often flawed, but the Doppler measurement is very accurate, a correct movement determination can be made. Compared to previous scan matching approaches, in which both the translation speed and the rotation are determined starting from the positions of the target points, the approach according to the invention can lead to more reliable results. Particularly in radar sensors, the measured Doppler velocity is often more precise than the measured position.FIG. 2 schematically illustrates a device 16 according to the invention. The device 16 comprises an input interface 20, a translation unit 22, a rotation unit 24 and an output unit 26. In particular, the device 16 comprises an object recognition unit 28, a classification unit 30 and an accumulation unit 32. The device 16 according to the invention can be integrated, for example, in a vehicle control device or can be embodied as part of a driver assistance system or can be implemented as a separate module. It is possible for the device 16 according to the invention to be partially or completely implemented in soft and / or in hardware. The various units and interfaces can be configured individually or in combination as a processor, processor module or software for a processor.Sensor data of an environment sensor are received via the input interface 20. In particular, a target point list of an FMCW radar sensor or of an FMCW lidar sensor is received. The list of target points comprises information about positions of target points on the object, which can be present in particular as angle information (azimuth and elevation) and a distance information (range). Furthermore, a Doppler velocity (radial velocity) is received for each of the plurality of target points. Typically, periodically updated target point lists generated based on scans of the environment sensor are received at successive points in time. The input interface 20 may be connected to a vehicle bus system, for example, in order to receive the data of the environment sensor. It is understood that data of several different environmental sensors can also be received via the input interface 20.In the translation unit 22, a translation speed of the object is first determined on the basis of the target point list. In particular, it is possible for destination point lists to be evaluated sequentially here, which have been generated or provided at successive sampling times. The translation speed is determined in particular by means of a state estimate and in the form of a Doppler profile of the object over an azimuth angle. The Doppler speeds of the various target points on the object are used directly in this respect in order to determine the translation speed.In the rotation unit 24, a rotation of the object is determined on the basis of the target point list on the one hand and the translation speed on the other hand. The determination of a rotation is understood in particular to mean the determination of a rotational speed. In this case, the previously determined translation speed can be included as a boundary condition. In contrast to previous scan matching approaches, in which a displacement and rotation were likewise determined, the translation speed is thus determined separately from the rotation in the approach according to the invention. During the rotation, it is assumed that the translation speed is already fixed. The translation speed is used as a boundary condition in this respect. In particular, it is then possible in principle to focus on minimizing a cost function in order to determine an appropriate rotation for the various target points in successive time steps. When determining the rotation, it is also possible to resort to modelling and state estimation.According to the invention, it is possible that the same or different sensor target lists are used when determining the translation speed as when determining the rotation. For example, sensor target lists of an FMCW radar sensor may be received and used to determine the translation speed. When determining the rotation, data of the same sensor can then be used on the one hand. On the other hand, it is also possible to rely on data from a lidar sensor or another radar sensor, since only positions (and no Doppler information) are necessary.The determined translation speed and rotation of the object are output via the output unit 26. In this case, further processing can be provided or the determined values can also be provided without further processing. In particular, a transformation matrix can be output via the output unit, which transformation matrix was determined on the basis of the translation speed and the rotation. The transformation matrix makes it possible to define the relative movement. The output unit 26 can in particular likewise be connected to a vehicle bus system of a vehicle, via which the ascertained data are forwarded to a vehicle control unit.In the object recognition unit 28, which is optionally present, it is recognized which target points are to be assigned to an object. In this respect, the data received via the input interface is analyzed in order to identify whether target points of an object or target points of a plurality of objects are contained in the target point list. The entire environment or the entire environment of the environment sensor (or of a vehicle) is usually mapped, so that the target point list comprises a plurality of objects. These objects are detected via the object detection unit 28, so that the target points of the target point list can be assigned to the objects. The translation unit and the rotation unit can then carry out the further processing of the data starting from target points which belong to exactly one object. Preferably, the processing of the object recognition unit 28 is provided directly after receiving the target point list.In the classification unit 30, which is also optionally provided, it can be established for an object whether it is a static or a dynamic object. Depending on whether the object is a static or a dynamic object, different parameterization can be carried out in the further processing in the translation unit 22 and in the rotation unit 24. For example, it may be advantageous if, in the case of a dynamic object, a higher accuracy is used in the detection of the relative movement (for example by means of more iteration steps or similar adaptations). Furthermore, in the case of a static object, it could be advantageous if all static objects are considered together, since the movement relative to the static objects is identical.In the accumulation unit 32, which is also optionally provided, an accumulated destination point list is generated. The accumulated goal point list includes goal points that were received at multiple times. In particular, a plurality of successive scans of an environment sensor can thus be combined. For this purpose, the target point lists or the scans, which are received at different times, must be transformed into a common coordinate system. Preferably, a sensor-fixed coordinate system is used, which is defined starting from the position of the environment sensor. The sensor targets received at different times must accordingly each be transformed into this coordinate system, since it is possible that the environment sensor (and the coordinate system) has moved further between two scans. For this transformation, the determined translation speed and the determined rotation can be used.FIGS. 3 and 4 schematically describe the determination of the translation speed by means of a Kalman filter. The representations are to be understood as the course of the Doppler velocity D over the azimuth angle A. The representations correspond to an observation of a relative speed in the azimuth range, since this is relevant in the case of an application for vehicles and a determination of a two-dimensional speed in a common subsurface plane. A speed on a road can therefore be determined, for example. The vertical component (elevation angle) is not considered here. According to the invention, however, a three-dimensional observation is naturally also possible.FIG. 3 shows a current (estimated) state x̂ k|k at a time step k. The state vector x includes a two-dimensional velocity v x, v y and a two-dimensional acceleration a x, a y. The system model further includes a noise term Q k, which is assumed to be normally distributed noise. Starting from the application of the system model, a new state estimate x̂ k|k+1, which is shown as a dashed line (prediction), results. The current measurement z k+1( v k+1) is represented by a dotted line and comprises a plurality of sensor target points Z1... ZN. For the sake of better clarity, only a few points are marked by way of example.FIG. 4 now shows that a gating method is used before the measurement update. The gating method is two-stage in the exemplary embodiment shown. First, a region is defined around the new state estimate (dash-dot lines), outside of which measurement points lying are discarded and not taken into account. In the example shown, points Z 3, Z 4 and Z 6 are thus discarded in this first step. These outlier points may represent, for example, target points in other objects or on the background that were incorrectly assigned to an object.In a further step, a RANSAC approach is applied to discard detections due to micro-doubler effects at wheels of a vehicle. In a first step, two sensor target points are randomly selected for this purpose. Then, statistics are established for each point based on a surrounding environment analysis and a point is discarded if it does not meet the corresponding condition. For example, the following systematics are used for this purpose:Here, r 1 and r 2 and ṙ 1 and ṙ 2 respectively denote the distance and radial speed of two randomly selected radar targets. Thus, two radar targets are selected for a defined number of iterations, from which a linear velocity hypothesis is calculated. For the calculated hypothesis, the number of targets that are within a certain speed band is counted. The hypothesis with the highest number of targets that are within the defined speed band is interpreted as a best estimate and used for discarding outlier targets. For example, points Z2 and Z5 are discarded.After the completion of the application of the gating method, only a few target points of the target point list remain in this respect (shown as continuous points, for example S1). By applying a measurement model, the current translation speed v x, v y is determined on the basis of the current measurements taking into account a noise term R k.As input for the Kalman filter, an estimate of the linear velocity is used, in which the mean quadratic error is minimized. For this purpose, the following equation system is established:Here, r i, ṙ i again denote the distance and radial speed of a particular radar target. From this equation system, the linear velocity v x, v y is resolved according toFIG. 5 schematically illustrates the approach according to the invention for determining the rotation. In particular, the rotation of the object is to be determined between the time steps k- 1 and k. Specifically, the rotational speed ω becomes.Starting from a cost functionShould a rigid body transformation g (consisting of a rotation and a translation) be determined which minimizes the mean square distance between corresponding points p i ∈ {p 1... p N} and q i ∈ {q 1... q M} from two successive radar target lists. The sought rigid body transformation can also be parameterized as a function of the linear and rotational speedHere, ω=[ω x ω y ω z]T denotes the angular velocity, and v=[v x v y v z]T denotes the linear velocity. The roof operator designates the obliquely symmetrical matrix shape of the respective vector. The linear velocity calculated beforehand can either be used directly in the optimization method. Alternatively, the Doppler velocity may be taken into account by an additional term in the cost function.Here, the vector of a radar target and ṙ i denotes the measured radial speed.Alternatively, the rotation may be determined by a particulate filter approach as shown in FIG. 5. For this purpose, a plurality of samples of possible rigid body transformations must be calculated and weighted with a suitable function. R and T denote the rotation matrix and translation vector of a sampled star body translation g, respectively.First, the linear velocity is determined as described above. Samples of the rotational speed ω (i) are then generated. Then, for each sample, the target points z i ∈{z 1... z N} of the previous time step are transformed and used as a basis (mean values) for a mixed density of the expected reflections in the current time stepHere, ξ denotes the velocity vector [v T ω T|T and Δt denotes the time difference between two successive radar measurements. denotes the reflections expected for the sample i. In a final step, the probabilities of a new target list or the position of the points of the target list are then analyzed based on the previous measurement step according toHere, π ij denotes the weighting of correspondence between targets from the different scans. The likelihood function for weighting the spatial correspondence of two targets is referred to as φ ij. A weighted distribution of sampled rotational speeds is produced.FIG. 6 schematically illustrates the data processing according to the invention using an example. In a first step P 1, a classification is made on the basis of received target lists (point clouds) as to whether it is a static or a dynamic object. Based on these classified detections, in a next step P2 a RANSAC approach is performed to estimate the translation speed as described above. Based on the estimated translation speed, Kalman filtering is applied in a third (optional) step P 3. There are two different approaches to estimating rotation below. On the one hand, in a matching approach P 4, a scan matching can be carried out based on a corresponding algorithm (for example Iterative Closest Point, ICP), in which the previously determined translation speed is introduced as a boundary condition. On the other hand, alternatively or additionally, in a step P4', a filtering can be carried out, in particular by means of a particle filter. Here, the previously determined translation speed can likewise be included as a boundary condition. Based on the determined rotation or angular velocity, it is then possible in a last step P 5 to accumulate the detection over a plurality of time steps in order to reconstruct an object.FIG. 7 schematically illustrates a method according to the invention for determining a relative movement between an environment sensor and an object. The method comprises steps of receiving S 10 a target point list of an environment sensor, determining S 12 a translation speed, determining S 14 a rotation and outputting S 16 the translation speed and the rotation. The method can be implemented in particular in software which is executed on a vehicle control device. It is likewise possible for the method to be carried out in an environment sensor, in particular in a radar and / or lidar sensor.The invention has been fully described and explained with reference to the drawings and the specification. The description and explanation are to be taken by way of example and not limitation. The invention is not limited to the disclosed embodiments. Other embodiments or variations will become apparent to those skilled in the art upon use of the present invention, as well as upon a detailed analysis of the drawings, disclosure and appended claims.In the claims, the words "comprise" and "with" do not exclude the presence of further elements or steps. The undefined article "a" or "an" does not exclude the presence of a plurality. A single element or unit may perform the functions of several of the units recited in the claims. An element, a unit, an interface, a device and a system can be partially or completely implemented in hardware and / or in software. The mere naming of some measures in several different dependent claims is not to be understood as meaning that a combination of these measures cannot likewise be used advantageously. A computer program can be stored / distributed on a non-volatile data carrier, for example on an optical memory or on a solid state drive (SSD). A computer program can be distributed together with hardware and / or as part of hardware, for example by means of the Internet or by means of wired or wireless communication systems. Reference signs in the patent claims should be understood to be non-limiting.Reference numerals denote reference numerals10 System 12 Environment sensor 14 Object 16 Device 18 Vehicle 20 Input interface 22 Translation unit 24 Rotation unit 26 Output unit 28 Object recognition unit 30 Classification unit 32 Accumulation unit

Claims

Device (16) for determining a relative movement between an environment sensor (12) and an object (14), comprising: an input interface (20) for receiving a target point list of the environment sensor with information about Doppler speeds and positions of target points on the object; a translation unit (22) for determining a translation speed of the object based on the target point list; a rotation unit (24) for determining a rotation of the object based on the target point list and the translation speed; and an output unit (26) for outputting the translation speed and the rotation of the object, wherein the translation unit is configured to determine the translation speed based on translation modelling over a plurality of time steps and a state estimate and / or the rotation unit is configured to determine the rotation based on rotation modelling over a plurality of time steps and a state estimate.The device (16) according to claim 1, wherein the rotation unit (24) is configured to determine the rotation with the inclusion of the translation speed as a boundary condition; and preferably to determine the rotation based on a minimization of a cost function.The device (16) according to any one of the preceding claims, wherein a state model of the rotation modelling does not depict an acceleration of the object (14); and / or a particle filter is applied for the state estimation in the rotation unit (24).The device (16) according to any one of the preceding claims, wherein the translation unit (22) is configured to determine the translation speed based on a random sample consensus, RANSAC, algorithm; and preferably to determine the translation speed in the form of a Doppler profile of the object (14) over an azimuth angle.The device (16) according to any one of the preceding claims, wherein a state model of the translation modelling maps an acceleration of the object (14); and / or a Kalman filter is applied for the state estimation in the translation unit (22).The apparatus (16) according to claim 5, wherein the translation unit (22) is configured to disregard outlier values in the target point list by applying a gating method.The device (16) according to any one of the preceding claims, comprising an object recognition unit (28) for recognizing target points belonging to an object (14), wherein the translation unit (22) is configured to determine the translation speed based on the target points belonging to the object; and the rotation unit (24) is configured to determine the rotation based on the target points belonging to the object.The device (16) according to claim 7, comprising a classification unit (30) for recognizing whether the object (14) is a static or a dynamic object based on the target point list, wherein the translation unit (22) is configured to determine the translation speed depending on whether the object is a static or a dynamic object.The device (16) according to any one of the preceding claims, comprising an accumulation unit (32) for generating an accumulated target point list with information about positions of target points on the object (14) based on target point lists received over a plurality of time steps and a transformation of the positions of the target points based on the determined translation speed and the determined rotation.The apparatus (16) according to any one of the preceding claims, wherein the input interface (20) is configured to receive the target point list from an FMCW radar sensor, which is preferably mounted on a vehicle.The device (16) according to any one of the preceding claims, wherein the input interface (20) is configured to receive a first target point list of a first environment sensor (12) with information about Doppler speeds and positions of target points on the object (14) and a second target point list of a second environment sensor with information about positions of target points on the object; the translation unit (22) is configured to determine the translation speed based on the first target point list; and the rotation unit (24) is configured to determine the rotation based on the second target point list.The device (16) according to any one of the preceding claims, wherein the input interface (20) is configured to receive the list of target points from an environment sensor (12) attached to a vehicle (18); the translation unit (22) is configured to determine the translation speed in a two-dimensional coordinate system fixed to the sensor, which is oriented parallel to a ground plane of the vehicle; and the rotation unit (24) is configured to determine the rotation in the same coordinate system.A system (10) for determining relative motion between an environment sensor (12) and an object (14), comprising: an apparatus (16) according to any preceding claim; and an environment sensor for detecting objects in an environment of a vehicle (18).Method for determining a relative movement between an environment sensor (12) and an object (14), comprising the steps of: receiving (S10) a target point list of the environment sensor with information about Doppler speeds and positions of target points on the object; determining (S12) a translation speed of the object based on the target point list; determining (S14) a rotation of the object based on the target point list and the translation speed; and outputting (S16) the translation speed and the rotation of the object, wherein the translation speed is determined based on a translation modelling over a plurality of time steps and a state estimation and / or the rotation is determined based on a rotation modelling over a plurality of time steps and a state estimation.A computer program product having program code for performing the steps of the method of claim 14, when the program code is executed on a computer.

Images