SIMULATION OF LIGHT DETECTION AND RANGING BEAMS

The present invention is related to a method of simulating a Light Detection And Ranging (LIDAR) beam. In particular, the present invention is related to simulating a LIDAR beam through ray-tracing.

LIDAR is an optical method for measuring distance and veloci ty as well as for measuring atmospheric parameters. Laser rays or beams of several rays are used for the measuring.

LIDAR is used as a measuring instrument for stationary and mobile speed monitoring. In the field of construction and surveying LIDAR is used for distance measuring. Further, in Automatic Driver Assistance Systems (ADAS) , and in autonomous vehicles, LIDAR is used for both measuring distance, e.g. to detect objects in the surroundings of a vehicle and to create a virtual map of said surroundings, and measuring velocity, e.g. to implement a collision warning. Moreover, in the field of aviation, LIDAR is also used for detecting turbulences and wind shear. Consequently, the measurements done with LIDAR must be reliable and precise because critical decisions are based thereon.

In order to improve the reliability and precision of a LIDAR system configuration, testing and validation is aided with simulations of the respective LIDAR system. Thereto, High- quality, physics based simulations of LIDAR systems are need ed. For such high-quality, physics based simulations it is necessary to accurately simulate a LIDAR beam, to produce da ta which is as close as possible to real data.

It is known to simulate ideal point clouds with a single sim ulated ray, e.g. by ray-tracing up to the nearest intersec tion point with a surface, whereby only mirror-reflection is considered such that only one distance per ray is computed. Here, the single ray or sub-beams of a beam represented by single rays are emitted from one point (one point source) . Such simulation returns only point data and neither the beam nor the resulting signal is simulated. Further it is known, to simulate a LIDAR beam with a 2D tree canopy in a 2D grid using a probability map to determine how the light propagates (light is propagated through probability) , whereby Lambertian or mirror reflection can be considered. Here, many rays are needed to sufficiently simulate the LIDAR beam and still the simulation is limited to 2D only. Thereby, each of the rays updates a signal of the beam only at a single time point. So, many iterations of the simulation of one beam are averaged together for each beam. However, no multiple interactions of the whole beam (several rays) with the environment (e.g.

scattering of a ray on several surfaces) and no energy dis tribution of the ray along its cross-section are considered. Interaction between two or more subsequent beams is not con sidered .

Angela Kim et al . disclose in "Monte Carlo techniques for di rect lighting calculations" a Monte Carlo model of laser propagation through a tree which allows simulation of full- waveform LIDAR signatures. To create a simulated waveform, multiple LIDAR pulses are propagated through the system, and the average response is output as the simulated waveform.

Each LIDAR pulse is defined as having a certain length, energy level, and beam divergence. The pulse is divided into multi ple small bins, with each bin being treated as a separate pulse having an initial energy, location, and propagation di rection determined by its location in the overall LIDAR pulse. Each pulse bin is propagated with a user-defined step size. At each step, the location of the LIDAR pulse is tracked, and the material interaction is determined probabil istically according to the material probability maps, and reflectance properties. When the pulse inter-acts with materi al having an absorbance greater than 0, the energy of the pulse is reduced by that amount. The remaining energy is al lowed to continue propagating. If the energy is transmitted, the angle of propagation is not changed. If the energy is reflected, the angle of propagation is updated. The reflection can be either Lambertian or specular. If the reflection is Lambertian, every angle of reflection between 0 and n radians from the material sur-face is equally likely, so a random an gle of reflection is chosen. In the specular case, the reflec tion is assumed to be a mirrored reflection from the plane of the material being impacted.

Consequently, there is a need for a method of simulating a LIDAR beam that simulates all aspects and interactions with the environment of a LIDAR beam as precise as possible.

These objectives are achieved by the Method according to claim 1 and the apparatus according to the further independ ent claim. Further refinements of the present invention are subject of the dependent claims.

According to a first aspect of the present invention a method of simulating a Light Detection And Ranging (LIDAR) beam of a LIDAR sensor in a scene comprises the following sequential steps :

a) Discretising a source area of the LIDAR beam into sepa rate origins .

b) Selecting rays based on the origins.

c) The method further comprises the following sequential steps for each ray:

cl) Finding an intersection of the current ray with an object of the scene through ray-tracing. c2) The method further comprises the following steps if an intersection is found:

Determining a reflected energy of the current ray that is reflected from the object at the intersection back to the LIDAR sensor.

Adding the reflected energy to a signal.

Determining a scattered ray based on the cur rent ray scattered from the object at the in tersection . Continuing with step cl) for the scattered

ray, wherein a lower threshold for the energy of the scattered ray down to which it is simu lated and/or a maximal number of reflections or bounces is predefined.

c3) Continuing with step cl) for the next ray.

d) Storing and/or outputting the added up signal.

According to a second aspect of the present invention a sys tem for simulating a LIDAR beam of a LIDAR sensor in a scene is configured for implementing the method according to the first aspect and comprises a processing unit and a storage. The processing unit has a discretising module, a selecting module and a ray-tracing module. The discretising module is configured for discretising a source area of the LIDAR beam into separate origins. The selecting module is configured for selecting rays based on the origins. The ray-tracing module is configured for:

Finding an intersection of the current ray with an ob ject of the scene through ray-tracing for each ray.

Determining a reflected energy of the current ray that is reflected from the object at the intersection back to the LIDAR sensor if an intersection is found.

Adding the reflected energy to a signal if an intersec tion is found.

Determining a scattered ray based on the current ray scattered from the object at the intersection if an in tersection is found.

Continuing with finding an intersection for the scat tered ray wherein a lower threshold for the energy of the scattered ray down to which it is simulated and/or a maximal number of reflections or bounces is predefined. Continuing with finding an intersection for the next ray .

The storage is configured for storing the added up signal or an output configured for outputting the added up signal. A LIDAR beam is characterised by several parameters: e.g. the beam cross-section, the beam divergence, the beam emission pattern and the beam pulse pattern. A LIDAR beam commonly used in automotive has a rectangular or square cross-section within which a laser or multiple lasers have a emission pat tern comprising horizontal or vertical rows/stripes or any other suitable pattern including meander shapes, circles, el lipses, spirals, a Gaussian configuration and the like. The beam may originate from a point source or several point sources defining a source area. From the point source or the source area the beam diverges with a certain divergence angle or divergence angles, which may be measured relative to the emission direction of the point source or source area, re spectively. If the LIDAR beam is emitted by a point source, the beam has a certain initial source area some distance away from the actual point source. In order to be able to use the method for single laser and multi laser LIDAR beam simulation the beam is always simulated starting from a source area.

In the first step a) the source area is discretised into sep arate origins. The source area may be discretised into

100.000 or less, preferably 10.000 or less and most prefera bly 4000 or less origins. Each origin can be the origin of one single ray of the LIDAR beam. Each ray has a certain emitting direction such that the LIDAR beam is simulated as a truncated pyramid of arbitrary cross-sectional shape like a frustum originating at the source area and diverging with a certain angle of divergence from the emission direction of the source area. The direction of each ray may be character ised by two divergence angles, Q and cp. The two divergence angles may each be within one of two perpendicular planes spanned by the emission direction and a respective radial di rection perpendicular to the emission direction.

In the second step b) rays are selected based on the origins. Each of the origins may emit one single ray of the LIDAR beam that is simulated as a truncated pyramid of arbitrary cross- sectional shape (e.g. frustum) . Thereby, each ray has a cer- tain direction along which the ray radiates from the source area. The number of rays, that is selected based on the re spective origins, may be deliberately predetermined. Conse quently, only a subset of the whole LIDAR beam is selected. However, increasing the number of rays also increases the precision of the simulation results. The rays may be selected according to a certain pattern or rule or randomly.

Each of the selected rays is simulated propagating through the scene or environment and interacting with the objects in the scene.

In the first step cl) for each of the rays an intersection of the current ray with an object in the scene is found by means of ray-tracing. Ray-tracing is a method to sample geometry in a virtual scene known from computer graphics. In computer graphics, ray-tracing is used to create an image by shooting rays from a camera and instantaneously (no propagation time) accumulate light on a sensor pixel. In the present method the selected rays (of the laser) are shot originating from the respective origin of the source area and connected to the LIDAR sensor, where light energy is accumulated in the (time varying) signal. In contrast to the ray-tracing used in com puter graphics, ray-tracing in the present invention does not ignore propagation time of the rays.

If an intersection is found for the current ray in step cl), then in the second step c2) for each of the rays the reflect ed energy of the (originally emitted) ray and its scattered ray are determined. At the intersection with an object the current ray interacts with the surface of the object. One part of the energy of the current ray is reflected back to wards the LIDAR sensor as the reflected energy. The amount of energy that is reflected back depends on a material and sur face of the object. The interactions with the objects in the scene may be calculated based on radiance rather than energy with the Bidirectional Scattering Distribution Function

(BSDF) . If the BSDF is used, the radiance is only converted into the unit of energy when the reflected energy is added to the signal. Apart from that all interactions between the rays and objects are calculated in the unit of radiance. Further, the remaining energy of the current ray is scattered into a different direction with a certain scattering angle that de pends on the material and surface of the object. The reflect ed energy is added to the signal, representing the signal of the real LIDAR sensor, and the scattered ray is continued to be simulated. So, instead of continuing with the next ray of the LIDAR beam, the scattered ray is ray-traced through the scene, by continuing with step cl) for the scattered ray. If an intersection of the scattered ray with an object in the scene is found during the continued ray-tracing, again the reflected energy is determined, this time based on the scat tered ray, and added to the signal. There may be a check whether the reflected energy can propagate from the intersec tion to the (virtual) LIDAR sensor without any occlusion. Thereto, via a shadow ray the visibility of the respective intersection from the view point of the (virtual) LIDAR sen sor is checked, wherein the shadow ray is cast from the (vir tual) LIDAR sensor to the respective intersection. The scat tered ray (which is the current ray in this iteration) is al so scattered such that a (twofold) scattered ray is continued to be simulated. Accordingly step cl) is executed for the newly (twofold) scattered ray. The ray-tracing for scattered rays continues as long as an intersection at the surface of an object in the scene occurs for the scattered rays. In or der to avoid tracing a ray to infinity, a lower threshold for the energy of the scattered ray down to which it is simulated and/or a maximal number of reflections/bounces (e.g. 10 re flections) can be predefined.

If no (more) intersection of a current ray (originally emit ted ray or scattered ray, respectively) occurs during ray tracing in step cl) then according to step c3) for each of the rays, the next ray is ray-traced in step cl) . Consequent ly, each of the selected rays is ray-traced and, if an inter section is found, scattered and, in turn, the resulting scat- tered ray is traced until no more intersection with an object in the scene occurs or optionally until the lower threshold for the energy of the ray or the maximal number of reflec tions is reached. At each intersection the reflected energy based on the current ray (originally emitted ray or scattered ray, respectively) is added to the signal of the LIDAR sen sor .

After all rays are simulated in steps cl) to c3) the result ing signal containing the summed up reflected energies is outputted to another method or system or a user. The signal can alternatively or additionally be stored in a storage de vice .

With the present method arbitrary 3D geometry can be simulat ed, where real intersection points are found, and multiple bounces of the rays of the beam (multi-bounce rays) are con sidered and the reflected energy of each ray and also of any of its resulting scattered rays are added to the signal. This enables a more precise simulation of the actual signal of the real LIDAR sensor because all interactions of the beam and its scattered rays with the environment (object in the scene) are simulated. Further, only one simulation per LIDAR beam is necessary to simulate every single possible interaction of the rays of the beam with the environment. Consequently,

LIDAR systems can be configured with better tolerance and be tested and validated with higher precision. This adds to se curity for example in autonomous vehicles and ADAS systems for modern vehicles.

According to a refinement of the present invention in step b) the rays are randomly selected.

Based on a randomly selected number of rays, which represent the whole LIDAR beam, the physics of the interactions can be simulated with Monte-Carlo Integration (randomised quadrature or Crude Monte-Carlo Integration) . This enables a precise and reliable simulation of the whole LIDAR beam based on only a few rays. Thus, the computing time and effort can be significantly reduced.

According to a refinement of the present invention in step b) the rays are randomly selected based on at least one random number and preferably based on two random numbers.

The selection of rays used for the simulation of the LIDAR beam can be effected with the help of one single random num ber. For example the origins of the discretised surface area can be consecutively numbered and the rays can be selected by randomly choosing numbers of the origins. Further, the selec tion of rays can be based on two random numbers. For example two random numbers each for one of the two deflection angles can be randomly chosen or the origins can be numbered accord ing to their position in a x direction (dim x) and a perpen dicular y direction (dim y) and the origins of the rays can be randomly chosen based on these two dimensions.

Selecting the rays based on one or two random numbers is a very efficient way of determining the set of rays used for simulation of the LIDAR beam.

According to a refinement of the present invention in step b) the rays are randomly selected by importance sampling, low discrepancy sampling patterns or deterministic sampling pat terns .

Importance sampling is a general technique for estimating properties of the LIDAR beam, while only having samples gen erated from a different distribution namely the selected rays than the distribution of interest namely the whole beam. Low- discrepancy sampling patterns are quasi-random or sub-random patterns, due to their use as a replacement of uniformly dis tributed random numbers for choosing the rays. In quasi- Monte-Carlo method the lower discrepancy of the low- discrepancy sampling patterns is an important advantage. De- spite the deterministic character of deterministic sampling patterns, the latter can be used instead of a random selec tion scheme and nevertheless produce a suitably distributed selection of rays just like the random schemes mentioned be fore .

By using importance sampling, low discrepancy sampling pat terns or deterministic sampling patterns the computing time and effort can be further reduced.

According to a refinement of the present invention in step b) at least two, preferably at least 100, more preferably at least 500 and most preferably at least 1000 rays are select ed .

According to a refinement of the present invention in step b) at most 100000, preferably at most 10000, more preferably at most 5000 and most preferably at most 2000 rays are selected.

The more rays are selected for simulating the whole LIDAR beam, the more precise and reliable are the simulation re sults, namely the signal.

With ray-tracing a limited number of rays as subset of the whole LIDAR beam, the computing time and effort can be sig nificantly reduced while maintaining a sufficiently high pre cision and reliability of the simulation results.

According to a refinement of the present invention step c2) further comprises the following steps if an intersection is found :

Calculating an intersection time point ti based on a radiation time point t _R of the cur rent ray and on a length of the current ray from its origin to the intersection Si.

Calculating a sensing time point t _s of sensing the reflected energy at the LIDAR sensor based on the intersection time point ti and on the length of a distance from the intersection to the LIDAR sensor.

In the step of adding the reflected energy, the reflected energy is added to the signal at the sensing time point t _s .

If an intersection of the current ray (originally emitted ray or scattered ray, respectively) is detected in step cl) the total length of the current ray from the respective origin to the intersection or the total length of the scattered ray from the respective origin to the last intersection (along all preceding intersections) Si is used together with the speed of light c to calculate the travel time of the respec tive ray. The respective intersection time point ti is then calculated based on the time point of radiation t _R , which is the time point, when the respective ray was emitted from the source area.

The sensing time point t _s when the reflected energy is sensed at the LIDAR sensor is calculated based on the intersection time point ti and the length between the intersection and the LIDAR sensor s _s , wherein the length of path along which the reflected energy travels from the intersection to the LIDAR sensor is also determined by ray-tracing. ss

t _s -t,+ ₇

Based on the determined sensing time point t _s the reflected energy is added to the signal at said sensing time point t _s .

Tracing the total length of the ray and the length of the path of the reflected energy and calculating the respective time points for the rays and the resulting signal based on the different lengths is very efficient and thus reduces the computing effort. According to a refinement of the present invention step of determining a scattered ray comprises reducing the current ray by the reflected energy.

In case an intersection with an object in the scene is found in step cl), the current ray or the scattered ray, respec tively is reduced by the determined reflected energy that is reflected back to the LIDAR sensor from the surface of the object. The interactions between the current ray (originally emitted ray or scattered ray, respectively) and the objects in the scene may be calculated using the BSDF, using the unit of radiance rather than the unit of energy. Thus, the current ray may be "continued" in the scattered direction reduced by the reflected energy.

This provides a way of tracing the rays with explicitly low implementation effort. Thus, the complexity and computing ef fort are further reduced.

According to a refinement of the present invention each ray represents a time varying energy distribution of a laser pulse contained in a vector of time bins and the reflected energy represents a time varying energy distribution corre sponding to the energy of the laser pulse reflected from the object at the intersection back over time contained in a vec tor of time bins.

The rays originating from the source area are simulated with a time varying energy distribution. The energy distribution is modelled as the time based energy distribution of a laser pulse. This energy distribution is contained in a vector of time bins, where every time bin contains a certain energy value of the ray. The time bins may have a size of about 1 ps [Picosecond] to about 1 ns [Nanosecond], preferably of about 1 ps, more preferably of about 10 ps and most preferably of about 100 ps . The reflected energy represents a time varying energy distribution. The energy distribution of the reflected energy corresponds to the energy (or radiance) of the laser pulse that is reflected from the surface object at the inter section. The reflected energy (distribution) is sent back to the (virtual) LIDAR sensor where it is sensed over time. Ad ditionally, there may be a check whether the path along which the reflected energy travels from the intersection to the (virtual) LIDAR sensor is not blocked or occluded by an ob ject or other occlusion. In other words, the visibility of the intersection from the view point of the (virtual) LIDAR sensor is determined with a shadow ray cast from the (virtu al) LIDAR sensor to the respective intersection. The energy distribution of the reflected energy is contained in a vector of time bins. The time bins may have a size of about 1 ps to about 1 ns, preferably of about 1 ps, more preferably of about 10 ps and most preferably of about 100 ps .

With the present refinement the energy distribution of the LIDAR beam with all its rays is considered in the simulation such that a full physics based simulation is enabled, where the LIDAR beam is fully simulated. Thus, arbitrary material properties can be used in the simulation and a full-waveform signal is computed. This enables a more precise simulation of the actual signal of the real LIDAR sensor because all inter actions of the beam and its scattered rays with the environ ment (objects in the scene) are simulated. Consequently,

LIDAR systems can be configured with better tolerance and be tested and validated with higher precision. This adds to se curity for example in autonomous vehicles and ADAS systems for modern vehicles.

According to a refinement of the present invention the signal is distributed over time represented by a vector of time bins. The vector has a length corresponding to a predeter mined sampling time T _s .

The rays have an energy that is represented as a time varying energy distribution which is contained in a vector of time bins. To be able to map the energy distributions of all the reflected energies which are accordingly represented by vec- tors of time bins the signal is represented by a vector with a length corresponding to the sampling time T _s . The sampling time T _s can be about 100 ns to 50000 ns and preferably about 500 ns to 2000 ns. The time bins may have a size of about 1 ps [Picosecond] to about 1 ns [Nanosecond], preferably of about 1 ps, more preferably of about 10 ps and most prefera bly of about 100 ps . The energy distributions of the reflect ed energies are sampled at the time point when they hit the LIDAR sensor and added bin-wise to the signal.

By sampling each energy distribution at the time point when the reflected energy is detected with the (virtual) LIDAR sensor a full physics based, full-waveform signal is calcu lated that precisely represents the real signal. Consequent ly, LIDAR systems can be configured with high precision and also be tested and validated against a highly accurate simu lated signal.

According to a refinement of the present invention the re flected energies of precedent beams are accounted for in the present beam.

Due to very late reflection of a ray or a (multiply) scat tered ray of a preceding beam its respective reflected energy may arrive at the (virtual) LIDAR sensor after the time point on which the simulation for the preceding beam ended. This "belated" reflected energy is not disregarded but accounted for in the signal of the simulation of the present beam.

With this accounting for reflected energies of preceding beams the simulation considers even "artefacts" from preced ing beams. Therefore, "auxiliary" noise that is commonly used to account for such artefacts can be omitted. Consequently, an even more precise simulation of the signal of the LIDAR sensor is provided.

The processing unit may be controlled by a program implement ing the method as described above. The present invention and its technical field are subsequent ly explained in further detail by an exemplary embodiment shown in the drawing. The exemplary embodiment only conduces better understanding of the present invention and in no case is to be construed as limiting for the scope of the present invention. Particularly, it is possible to extract aspects of the subject-matter described in the figure and to combine it with other components and findings of the present description or figure, if not explicitly described differently. Equal reference signs refer to the same objects, such that explana tions from other figures may be supplementally used.

Fig. 1 schematically shows a flow chart of a method ac

cording to the first aspect of the present inven tion.

Fig . 2 schematically shows a system according to the sec ond aspect of the present invention.

Fig . 3 schematically shows a LIDAR beam.

Fig . schematically shows a vector of a reflected energy and a vector of a signal of a LIDAR sensor,

Fig . schematically shows multi-bounce ray-tracing.

In Fig. 1 a flow chart of a method 10 of simulating a Light Detection And Ranging (LIDAR) beam of a LIDAR sensor in a scene is schematically depicted. The method comprises a first step a) of discretising 11 a source area, a subsequent second step b) of selecting 12 rays, a subsequent third iteration step c) for ray-tracing 13 and a fourth step of storing and/or outputting 14. The third iteration step c) for ray tracing 13 comprises a first step cl) of finding 15 an inter section, a subsequent second iteration step c2) for signal calculation 16 and a third step c3) of continuing 17 with step cl) . The second iteration step c2) for signal calcula tion comprises a step of determining 16.1 a reflected energy, a step of calculating 16.2 an intersection time point ti, a step of calculating 16.3 a sensing time point t _s , a step of adding 16.4 the reflected energy, a step of determining 16.5 a scattered ray and a step of continuing 16.6 with step cl) .

The source area of the LIDAR beam is discretised into origins in the first step a) of discretising 11. Each origin is the basis from which one single ray of the LIADR beam is emitted in a specific direction. The direction is given by two diver gence angles f and Q measured against the emission direction on the source area.

Based on one or two random numbers 2 to 100000, preferably 100 to 10000, more preferably 500 to 5000 and most preferably 1000 to 2000 rays are selected in the second step b) of se lecting 12. The selecting 12 is based on importance sampling, low discrepancy sampling patterns or deterministic sampling patterns. Each ray represents a time varying energy distribu tion of a laser pulse. The energy distribution of the laser pulse is simulated as a normal distribution contained in a vector of time bins. The time bins have a size of 100 ps [Pi coseconds] . The vector of the ray has a total length of 10 ns [Nanoseconds] containing 100 bins.

In the following third iteration step c) for ray-tracing 13 each of the (originally emitted) rays or scattered rays (see below) is traced and an intersection with an object in a vir tual scene is searched according to the first step cl) of finding 15. If an intersection is found the second iteration step c2) for signal calculation 16 is executed.

Based on the found intersection with a surface of the object a reflected energy is determined based on the material and surface of the object in the step of determining 16.1 for the current ray or the scattered ray (see below) . The reflected energy represents a time varying energy distribution of the reflected energy corresponding to the energy of the laser pulse that is reflected from the surface object at the inter section. The reflected energy (distribution) is sent back to the (virtual) LIDAR sensor where it is sensed over time. The energy distribution of the reflected energy is contained in a vector of time bins. The time bins have a size of 100 ps and the vector of the reflected energy has a total length of 10 ns containing 100 bins. The reflected energy is added to the signal representing the signal of the real LIDAR sensor.

Further an intersection time point ti is calculated in the step of calculating 16.2 based on a radiation time point t _R of the respective ray, which is the time point, when the re spective ray was emitted from the source area, and a total length of the current ray from the respective origin to the intersection or the total length of a scattered ray (see be low) from the respective origin to the last intersection (along all preceding intersections) Si together with the speed of light c.

Further, in the step of calculating 16.3 a sensing time point t _s , giving the time point when the reflected energy is sensed at the LIDAR sensor, is calculated based on the intersection time point ti and the length between the intersection and the LIDAR sensor s _s , wherein the length of path along which the reflected energy travels from the intersection to the LIDAR sensor is also determined by ray-tracing. ss

t _s -t,+ ₇

Based on the determined sensing time point t _s the vector of the reflected energy (100 bins of 100 ps each) is added in the step of adding 16.4 to the signal at said sensing time point ts· The signal is distributed over time represented by a vector of time bins of 100 ps . The vector has a total length corresponding to a sampling time T _s of 700 ns and con taining 7000 bins. Additionally, in the step of determining 16.5 a scattered ray is determined based on the surface and material of the object at the intersection. The scattered ray has an energy distri bution according to the ray but reduced by the scattered en ergy. The scattered ray leaves the intersection under a scat tering angle depending on the surface and material of the ob ject at the intersection.

If an intersection was found in step cl) of finding 15 the method 10 is continued in the step of continuing 16.6 at step cl) of finding 15 with the scattered ray as current ray.

If no intersection is found in step cl) of finding 15, then the method is continued in the step c3) of continuing 17 at step cl) of finding 15 with the next ray as current ray until all rays are processed.

After all rays have been processed the method is continued at the step d) of storing or outputting 14, where the added up signal is output to another method and/or system and/or to a user and/or stored in a storage.

Additionally to the reflected energies coming from the (orig inally emitted) rays and scattered rays of the present beam the method 10 may also consider reflected energies from

(originally emitted) rays or scattered rays of preceding beams in the step of adding 16.5, when the reflected energy arrives at the (virtual) LIDAR sensor after the sampling time T _s . Thus, reflected energies that are not considered in the signal of the beam that effected the reflected energies be cause they arrive "belated" after the sampling time T _s can be considered in the signal of the next beam.

In Fig. 2 a system 20 for simulating a LIDAR beam of a LIDAR sensor in a scene is schematically depicted. The system 20 comprises a processing unit 21 and a storage and/or output 22. The processing unit has a discretising module 23 config ured for discretising a source area of the LIDAR beam into separate origins, a selecting module 24 configured for se- lecting rays based on the origins, wherein each ray repre sents a time varying energy distribution of a laser pulse contained in a vector of time bins, and a ray-tracing module 25. The ray-tracing module 25 is configured for finding an intersection of the current ray with an object of the scene through ray-tracing for each ray. The ray-tracing module is further configured for determining a reflected energy of the current ray that is reflected from the object at the inter section back to the LIDAR sensor if an intersection is found, wherein the reflected energy represents a time varying energy distribution corresponding to the energy of the laser pulse reflected from the object at the intersection back over time contained in a vector of time bins. Also the ray-tracing mod ule 25 is configured for calculating an intersection time point ti based on a radiation time point t _R of the current ray and on a length of the current ray from its origin to the intersection Si and for calculating a sensing time point t _s of sensing the reflected energy at the LIDAR sensor based on the intersection time point ti and on the length of a dis tance from the intersection to the LIDAR sensor s _s , wherein in the step of adding the reflected energy, the reflected en ergy is added to the signal at the sensing time point t _s . Ad ditionally, the ray-tracing module 25 is configured for add ing the reflected energy to a signal if an intersection is found. Moreover, the ray-tracing module 25 is configured for determining a scattered ray based on the current ray scat tered from the object at the intersection if an intersection is found and for reducing the current ray by the reflected energy as well as for continuing with finding an intersection of the current ray for the scattered ray. Lastly, the ray tracing module is also configured for continuing with finding an intersection of the current ray for the next ray. The storage 22 is configured for storing the added up signal and the output 22 configured for outputting the added up signal.

The processing unit 21 can be a Central Processing Unit (CPU) or Graphics Processing Unit (GPU) like a Microcontroller (mq , an integrated circuit, an Application-Specific Inte- grated Circuit (ASIC) , an Application-Specific Standard Prod uct (ASSP) , a Digital Signal Processor (DSP) , a Field Pro grammable Gate Array (FPGA) and the like. The processing unit 21 may be controlled by a program implementing the method of Fig. 1 or as described above. The modules 23 to 25 may be part of the processing unit 21 or implemented on the pro cessing unit 21 or separate units in communicative connection with the processing unit 21. The storage 22 can be a data storage like a magnetic storage/memory (e.g. magnetic-core memory, magnetic tape, magnetic card, magnet strip, magnet bubble storage, drum storage, hard disc drive, floppy disc or removable storage), an optical storage/memory (e.g. holo graphic memory, optical tape, Tesa tape, Laserdisc, Phase- writer (Phasewriter Dual, PD) , Compact Disc (CD) , Digital Video Disc (DVD) , High Definition DVD (HD DVD) , Blu-ray Disc (BD) or Ultra Density Optical (UDO) ) , a magneto-optical stor age/memory (e.g. MiniDisc or Magneto-Optical Disk (MO-Disk)), a volatile semiconductor/solid state memory (e.g. Random Ac cess Memory (RAM) , Dynamic RAM (DRAM) or Static RAM (SRAM) ) , a non-volatile semiconductor/solid state memory (e.g. Read Only Memory (ROM) , Programmable ROM (PROM) , Erasable PROM (EPROM), Electrically EPROM (EEPROM) , Flash-EEPROM (e.g. USB- Stick) , Ferroelectric RAM (FRAM) , Magnetoresistive RAM (MRAM) or Phase-change RAM) or a data carrier/medium. The output 22 may be any kind of port or link or interface capable of com municating information to another (e.g. WLAN, Bluetooth, ZigBee, Profibus, ETHERNET etc.) system or to a user (Dis play, Printer, Speaker etc.).

In Fig. 3 a LIDAR beam 30 is schematically depicted. The LIDAR beam 30 has a rectangular source area 31 with a first dimension dim X and a second dimension dim Y. Here the source area 31 is a square with dim X = dim Y. The LIDAR beam is made up by three striped emission pattern of a single laser diode scanning the environment or scene. The LIDAR beam 30 has two divergence angles f and Q, wherein here f = Q. Conse quently, the LIDAR beam 30 has the shape of a frustum 32. The LIDAR beam 30 is made up by rays 33 originating from the source area 31 and each projecting from one origin of the source area 31 in a unique direction defined by f and Q.

In Fig. 4 a vector of a reflected energy 41 and a vector of a signal 42 of a LIDAR sensor are schematically depicted, each ray of the LIDAR beam may be represented by an energy distri bution. This energy distribution corresponds to an energy distribution of a laser pulse of a LIDAR sensor. The energy distribution of the laser pulse is simulated as a normal dis tribution. As the ray is scattered at an intersection with an object a part of its energy is reflected back to the LIDAR sensor as reflected energy. The energy distribution 40 of the reflected energy is also a normal distribution. At the sens ing time point t _s the reflected energy is added to the signal simply by adding the bins 41. i, i = 1..100, of the vector 41 of the reflected energy to the respective bins 42. j, j = 1..7000, of the vector 42 of the signal.

In Fig. 5 multi-bounce ray-tracing is schematically depicted. A ray 32 originating from the source area 31 is scattered at an object 50 one part of its energy is reflected back to the LIDAR sensor as reflected energy 52. The ray is also contin ued as scattered ray 51 reduced by the reflected energy 52. The scattered ray is scattered at another object 50 and part of the energy of the scattered ray 51 is reflected back to the LIDAR sensor as another reflected energy 53. This multi bounce ray-tracing may be executed for any arbitrary number of scatterings of one ray 33.

Although specific embodiments have been illustrated and de scribed herein, it will be appreciated by those of ordinary skill in the art that a variety of alternate and/or equiva lent implementations exist. It should be appreciated that the exemplary embodiment or exemplary embodiments are only exam ples, and are not intended to limit the scope, applicability, or configuration in any way. Rather, the foregoing summary and detailed description will provide those skilled in the art with a convenient road map for implementing at least one exemplary embodiment, it being understood that various chang es may be made in the function and arrangement of elements described in an exemplary embodiment without departing from the scope as set forth in the appended claims and their legal equivalents. Generally, this application is intended to cover any adaptations or variations of the specific embodiments discussed herein.

In the foregoing detailed description, various features are grouped together in one or more examples for the purpose of streamlining the disclosure. It is understood that the above description is intended to be illustrative, and not restric tive. It is intended to cover all alternatives, modifications and equivalents as may be included within the scope of the invention. Many other examples will be apparent to one skilled in the art upon reviewing the above specification.

Specific nomenclature used in the foregoing specification is used to provide a thorough understanding of the invention. However, it will be apparent to one skilled in the art in light of the specification provided herein that the specific details are not required in order to practice the invention. Thus, the foregoing descriptions of specific embodiments of the present invention are presented for purposes of illustra tion and description. They are not intended to be exhaustive or to limit the invention to the precise forms disclosed; ob viously many modifications and variations are possible in view of the above teachings. The embodiments were chosen and described in order to best explain the principles of the in vention and its practical applications, to thereby enable others skilled in the art to best utilize the invention and various embodiments with various modifications as are suited to the particular use contemplated. Throughout the specifica tion, the terms "including" and "in which" are used as the plain-English equivalents of the respective terms "compris ing" and "wherein," respectively. Moreover, the terms

"first," "second," and "third," etc., are used merely as la bels, and are not intended to impose numerical requirements on or to establish a certain ranking of importance of their objects. In the context of the present description and claims the conjunction "or" is to be understood as including

("and/or") and not exclusive ("either ... or") .