SE1951488A1

SE1951488A1 - Method and system for predicting movement

Info

Publication number: SE1951488A1
Application number: SE1951488A
Authority: SE
Inventors: Carl-Fredrik Alveklint; Simon Vajedi; Thomas Klintberg
Original assignee: Forsete Group Ab
Priority date: 2019-12-18
Filing date: 2019-12-18
Publication date: 2021-06-19
Also published as: WO2021126062A1

Abstract

The present disclosure generally relates to a novel concept for using machine learning to predict a future movement associated with a first movable object. The present disclosure also relates to a corresponding system and a computer program product.

Description

METHOD AND SYSTEM FOR PREDICTING MOVEMENT TECHNICAL FIELDThe present disclosure generally relates to a novel concept to use machineleaming for predicting a ﬁature movement associated with a first movable object. The present disclosure also relates to a corresponding system and a computer program product.

BACKGROUND Recent advances in computers and communications have had great impact onhow to predict the likelihood of an event occurring within a certain amount of time or theamount of time until an event is likely to occur. Predictive modeling generally refers totechniques for extracting information from data to build a model that can predict an outputfrom a given input. Predicting an output can include predicting future trends or behaviorpattems, to name a few examples.

In some implementations machine leaming is used in relation to suchpredictive modeling. Machine leaming is a form of artificial intelligence that is employed toallow computers to evolve behaviors based on empirical data. Machine leaming may takeadvantage of training examples to capture characteristics of interest of their unknownunderlying probability distribution. Training data may be seen as examples that illustraterelations between observed variables. Specifically, focus is nowadays towards ways toautomatically leam to recognize complex pattems and make intelligent decisions based ondata.

One possible application of such machine leaming based prediction modellingis in relation to semi or fully autonomous vehicles. An autonomous vehicle is typicallyequipped a computer system implementing the prediction modelling based on informationfrom a plurality of sensors, such as cameras, radar, and other similar devices, that is used tointerpret a surrounding of the vehicle. The computer system is adapted to executes numerousdecisions while the autonomous vehicle is in motion, such as speeding up, slowing down,stopping, tuming, etc. Autonomous vehicles may also use the cameras, sensors, and globalpositioning devices to gather and interpret images and sensor data about its surroundingenvironment, e.g., pedestrians, bicyclists, other vehicles, parked cars, trees, buildings, etc.

An exemplary implementation of such a computer system is disclosed inUS9248834, presenting a solution where information from the computer system is used to form a detailed map about the vehicle°s surrounding to allow the vehicle to safely maneuver the vehicle in various environments. This detailed map may describe expected conditions ofthe vehicle's environment such as the shape and location of roads, parking spots, dead zones,traffic signals, and other objects. In this regard, the detailed map may be used to assist inmaking driving decisions involving intersections and traffic signals, Without the interaction ofa driver/operator.

Even though US9248834 provides an interesting approach to apply aprediction scheme for improving an overall safety in operating an autonomous vehicle, thereis alWays room for further improvements and expansion of such technology, With theintention to reduce computational complexity While at the same time improving accuracy of the prediction.

SUMMARY In view of above-mentioned and other drawbacks of general prior art Withinthe technical area, it is an object of the present disclosure to provide improvements in relationto prediction of future movement associated With movable objects.

According to an aspect of the present disclosure, it is therefore provided acomputer implemented method for using machine leaming to predict a future movementassociated With a first movable object, the method comprising the steps of receiving asequence of representations indicative of present behavior of the first movable object,Wherein the sequence of representations comprises a plurality of different behavioral types,forming a plurality of behavioral maps for the first movable object by transforrning theplurality of representations for each of the plurality of different behavioral types using at leastone machine leaming component, and predicting the future movement of the first movableobject based on the plurality of behavioral maps. The steps in line With the present disclo suremay preferably be performed using processing circuitry comprised With a computer system.

In accordance to the present disclosure, the present behavior of a first object isanalyzed to determine a likely future "action" of the first object. In line With the presentdisclosure, this is achieved by separating the received sequence of representations, forexample being data received form an image sensor, a lidar sensor, a radar sensor, etc., intodifferent behavioral types. The different behavior of the object may generally be dependenton the type of object, Whereby e. g. in case the first object is a human it Will have differentbehavior as compared to an animal or a movable machine/robot.

The data relating to the different behaviors may then in line With the present disclosure be handled in parallel, using at least one machine leaming component for forrning behavioral maps for each of the plurality of different behavioral types. The behavioral mapsmay in tum be seen as how the first object presently behave within its present context anddepending on the specific behavioral type.

Once the behavioral maps have been deterrnined, they are possibly combined(or fused together) for predicting the future movement of the first movable object.Accordingly, in line with the present disclosure the overall behavior of the first object isobserved and analyzed, possibly segmented, to then be combined again for predicting thefuture movement of the first object.

Preferably, the method according to the present disclosure further comprise thestep of filtering at least a portion of the plurality of behavioral maps using a predeterrninedfiltering scheme. Such a filtering scheme may for example be based on the behavioral type ora sensor used for acquiring the sequence of representations. That is, different type of filteringschemes may be used for different sensor, with the intention to reduce an amount of noisebeing introduced when deterrnining the plurality of behavioral maps.

Advantageously, the behavioral types are selected from a group comprisingobject pose, head pose, eye gaze, velocity, position, acceleration or an interaction map for thefirst movable object. The expression "interaction map" should within the context of thepresent disclo sure be understood to relate to how the first object relates to other first objects(such as when the first object is part of a "crowd") and in relation to fixed objects in thevicinity of the first object. Accordingly, it may in some embodiments be desirable to allowthe sequence of representations to comprise information relating to at least one static objectlocated in a vicinity of the first movable object.

In some embodiments of the present disclosure a dedicated machine leamingcomponent is selected for each of at least a portion of the behavioral types. Such animplementation may for example make a formation of the dedicated machine leamingcomponent slightly simpler as such a component may be directed only to a specificbehavioral type. Such an implementation may also allow for a modularity of theimplementation, possibly allowing for a ﬂexible introduction of ﬁirther (and updated)dedicated machine leaming components over time. That said, within the context of thepresent disclo sure it may also be possible to form a single machine leaming component thathas been formed to handle all of the different behavioral types handled in line with thepresent disclo sure.

In deterrnining the behavioral map, the at least one machine leaming component reviews at least a present and a predeterrnined number of previous representations for the first movable object, where the predeterrnined number of previous representations forthe first movable object in some embodiments may be dynamic. Accordingly, the machineleaming component will in such an embodiment not just review the present behavior of thefirst object, but also take into account how the object just previously behaved. It may in linewith the present disclo sure be possible to apply different weights to the differentrepresentations, where representations in the past generally will be provided with a lowerweight as compared to newly received representations.

In preferred embodiments of the present disclosure, the scheme according to thepresent disclo sure is performed using processing circuitry comprised with a second object,the second object being different from the first object. The second object may be stationary ormovable. For example, a stationary second object may be a security camera, a traffic light,etc. Correspondingly, a movable second object may generally be any form of craft or vessel.Possible movable second objects also include any form of unmanned aerial vehicles (UAV)(such as drones) or vehicles (such for example as semi or ﬁally autonomous vehicles). Furtherboth present and future stationary or movable second objects are possible and within thescope of the present disclosure.

When the second object is movable, it may be possible to adapt the schemeaccording to the present disclosure to also include the step of forming control commands forthe second movable object based on the predicted future movement of the first movableobject, for example represent a trajectory for the second movable object. As an altemative,the control commands could be direct actuation commands for controlling an operation of thesecond object.

Preferably, the control commands for the second movable object are formed toreduce an interaction between the first and the second object. In a general implementation,such as when the second object is a vehicle and the first object is a human, the overallintention may for example be to ensure that the vehicle is controlled in a way to ensure thatthe human stays safe without being hit by the car.

Along the same line, in some embodiment of the present disclo sure, as willelaborated below in relation to the detailed description, the second object may be an airbomedrone on a recognizance mission with the intention of "not being seen" by the first object,where again the first object may be a human, for example being part of a plurality of humansarranged in a crowd. In such an embodiment it will typically be desirable to analyze the human(s) in regards to their movements/head pose/ eye gaze, etc., and to form the control commands for the drone such that the drone is arranged to ﬂy outside of e.g. a line of sightfor the human(s).

It should further be understood that the scheme according to the presentdisclosure may be adapted in a manner to increase the interaction between the first and thesecond object, such as in relation to implementations Where it is desirable to arrange the firstand the second object in close vicinity of each other. Such embodiments may for examplefind its Way into industrial applications Where a robot being the second object is to interactWith an animal being the first object.

According to an aspect of the present disclosure, there is further provided acomputer system comprising processing circuitry, the computer system arranged to predict afuture movement associated With a first movable object by adapting the processing circuitryto receive a sequence of representations indicative of present behavior of the first movableobject, Wherein the sequence of representations comprises a plurality of different behavioraltypes, form a plurality of behavioral maps for the first movable object by transforrning theplurality of representations for each of the plurality of different behavioral types using at leastone machine leaming component, and predict the future movement of the first movableobject based on the plurality of behavioral maps. This aspect of the present disclo sureprovides similar advantages as discussed above in relation to the previous aspects of thepresent disclo sure.

As indicated above, the computer system may in some embodiments becomprised as an onboard component of a second object (being different from the first object).The second object may as such be movable or stationary.

According to a further aspect of the present disclosure, there is provided acomputer program product comprising a non-transitory computer readable medium havingstored thereon computer program means for operating a computer system to predict a futuremovement associated With a first movable object using machine leaming, the computersystem comprising processing circuitry, Wherein the computer program product comprisescode for receiving, at the processing circuitry, a sequence of representations indicative ofpresent behavior of the first movable object, Wherein the sequence of representationscomprises a plurality of different behavioral types, code for forrning, at the processingcircuitry, a plurality of behavioral maps for the first movable object by transforrning theplurality of representations for each of the plurality of different behavioral types using at leastone machine leaming component, and code for predicting, at the processing circuitry, the future movement of the first movable object based on the plurality of behavioral maps. Also this aspect of the present disclo sure provides similar advantages as discussed above inrelation to the previous aspects of the present disclosure.

A software executed by the server for operation in accordance to the presentdisclosure may be stored on a computer readable medium, being any type of memory device,including one of a removable nonvolatile random access memory, a hard disk drive, a ﬂoppydisk, a CD-ROM, a DVD-ROM, a USB memory, an SD memory card, or a similar computerreadable medium known in the art.

In summary, the present disclosure generally relates to a novel concept forusing machine leaming to predict a future movement associated with a first movable object.The present disclosure also relates to a corresponding system and a computer programproduct.

Further features of, and advantages with, the present disclosure will becomeapparent when studying the appended claims and the following description. The skilledaddressee realize that different features of the present disclosure may be combined to createembodiments other than those described in the following, without departing from the scope of the present disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS The various aspects of the present disclosure, including its particular featuresand advantages, will be readily understood from the following detailed description and theaccompanying drawings, in which: Fig. l conceptually illustrates a computer system according to an embodimentof the present disclosure connected to a vehicle; Figs. 2A - 2C presents a possible use of the computer system in a relation toan airbome drone; Fig. 3 is a ﬂow chart illustrating the steps of performing the method accordingto a currently preferred embodiment of the present disclosure, and Fig. 4 conceptually shows a possible implementation of a machine leaming component that may be used on relation to the present disclosure.

DETAILED DESCRIPTIONThe present disclosure will now be described more fully hereinafter withreference to the accompanying drawings, in which currently preferred embodiments of the present disclo sure are shown. This present disclosure may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein;rather, these embodiments are provided for thoroughness and completeness, and fully conveythe scope of the present disclosure to the skilled person. Like reference characters refer tolike elements throughout. The following examples illustrate the present disclosure and are notintended to limit the same.

Tuming now to the drawings and to Fig. 1 in particular, there is conceptuallyillustrated a computer system 100 according to an embodiment of the present disclosure. Thepurpose of the computer system 100 is, in one embodiment, to dynamically observe andanalyze a behavior of a first movable object for predicting the future movement of the firstobject.

In a possible embodiment, the computer system 100 comprises processingcircuitry 102 and a plurality of sensors, for example including a camera 104, a lidar sensor106 and/or a radar sensor 108.

For reference, the processing circuitry 102 may for example be manifested as ageneral-purpose processor, an application specific processor, a circuit containing processingcomponents, a group of distributed processing components, a group of distributed computersconfigured for processing, a field programmable gate array (FPGA), etc. The processor maybe or include any number of hardware components for conducting data or signal processingor for executing computer code stored in memory. The memory may be one or more devicesfor storing data and/or computer code for completing or facilitating the various methodsdescribed in the present description. The memory may include volatile memory or non-volatile memory. The memory may include database components, object code components,script components, or any other type of information structure for supporting the variousactivities of the present description. According to an exemplary embodiment, any distributedor local memory device may be utilized with the systems and methods of this description.According to an exemplary embodiment the memory is communicably connected to theprocessor (e. g., via a circuit or any other wired, wireless, or network connection) andincludes computer code for executing one or more processes described herein.

Preferably, the computer system 100 is connected to a network, such as theIntemet 110, allowing the computer system 100 to communicate and exchange informationwith e. g. a remotely located server 112, having a thereto connected remote database 114. Theremotely located server 112 may be arranged to receive information about the first object and to provide the computer system 100 with general directions for operation.

To allow the computer system 100 to communicate with the remotely locatedserver 112, the computer system 100 may further comprise a transceiver 116 adapted to allowfor any form of wireless connections like WLAN, CDMA, GSM, GPRS, 3/4/5G mobilecommunications, or similar. Other present of future wireless communication protocols arepossible and within the scope of the present disclosure.

With further reference to Figs. 2A - 2C, there is shown a possible approach ofimplementing the computer system 100 in relation to a movable second obj ect, where in Fig.2 the movable second object is in the form of an airbome drone 200. As indicated above, thescheme according to the present disclo sure could be implemented in relation to any otherform of stationary or movable (second) objects for deterrnining a movement of another (first)object (or e.g. group of objects).

As shown in Fig. 2A, a group of persons 202, 204, 206, 207 are illustrated aswalking along a road 208. Each of the persons 202, 204, 206, 207 are within the scope of thepresent disclosure each defined as a first moving object. The road 208 as well as e.g. the trees210 along the road 208 (as well as other similar objects) are considered to be stationaryobjects in a vicinity of the first moving object(s) 202, 204, 206, 207.

Furthermore, in Fig. 2A there is shows a plurality of airbome drones 212, 214,216 flying at a distance from the persons 202, 204, 206, 207. Each of the drones comprises acontrol system 100 as presented in Fig. 1 and further detailed in Fig. 2B, preferably eacharranged in communication with the remote server 114. It may also be possible to allow theairbome drones 212, 214, 216 to communicate directly between each other, using any formof wireless communication protocol, e. g. as suggested above.

During operation of the control system 100 when comprised with e. g. theairbome drone 212 and with further reference to Fig. 3, the airbome drone 212 (or group ofairbome drones) has been assigned to perform a recognizance mission, dynamicallytravelling from a start position to a destination. The destination could be the same position asthe start position. When performing the recognizance mission, it is desirable to minimize anyinteractions with e. g. humans (such as persons 202, 204, 206, 207), since such an interactionpotentially could result in an undesirable knowledge of the fact that the recognizance missionis perforrned/ongoing.

To minimize the interaction between the airbome drone 212 and the persons202, 204, 206, 207, the control system 100 has been adapted to implement the scheme according to the present disclosure. As exemplified in Fig. 2B, the camera 104 (or any other sensor comprised with the drone 212) is used for collecting information as to a surroundingofthe drone 212.

The inforrnation from the camera 104 is received, S 1, at the processingcircuitry 102. When the drone 212 is within a distance from the persons 202, 204, 206, 207allowing sufficiently clear images of the person 202, 204, 206, 207 to be collected, then theimages may be defined to comprise representations indicative of a present behavior of thepersons 202, 204, 206, 207. Such behavior may be defined to include a plurality of differentbehavioral types, where the plurality of different behavioral types for example may include adirection of movement of the persons 202, 204, 206, 207, sizes of the persons 202, 204, 206,207. Other behavioral types that may be identified from e. g. the images from the camera 104may include different positions of the persons 202, 204, 206, 207 face, head, or upper body.These positions may, for example, be the eyes, eye-lids, eyebrows, nose, mouth, cheek, neck,shoulders, arms, etc.

The camera 104 may also detect, with ﬁirther reference to Fig. 2C, if the head,or eyes, of the operator is rotating to the right or left (yaw), 218, rotating up or down (pitch),220, or, in the case of the head movements, leaning towards the right or left shoulder (roll),222. If providing the camera 104 with e.g. high-quality optics, it could also be possible todetect e.g. an eye gaze direction for each of the persons 202, 204, 206, 207.

Based on the received images from the camera 104 (or from any other sensor),it may be possible for the processing circuitry 102 to form, S2, a plurality of behavioral mapsfor the first movable object by transforrning the plurality of representations for each of theplurality of different behavioral types using at least one machine leaming component.

As indicated above, in deterrnining the behavioral map, the at least onemachine leaming component reviews at least a present and a predeterrnined number ofprevious representations for the first movable object, i.e. possible historical data as to howthe persons 202, 204, 206, 207 have behaved both seen separately and together.

Once the behavioral maps have been deterrnined, then the processing circuitry102 may combine and/or fuse together information comprised with the behavioral maps forpredicting, S3, the future movement of the persons 202, 204, 206, 207. That is, each of thebehavioral maps will provide a component in the overall prediction of how each of thepersons 202, 204, 206, 207 seen individually as well as the group of persons 202, 204, 206,207 likely will behave, and specifically move, including for example in what direction the 202, 204, 206, 207 likely will look, how fast they will walk, etc.

Based on the predicted future movement of the persons 202, 204, 206, 207 theprocessing circuitry 102 may determine or form control commands that may be used by thedrone 212 (as well as the other drones 214, 216) to operate the drone 212 such that it will bepositioned in an undisclo sed position in relation to the predicted future movement of thepersons 202, 204, 206, 207. The control commands may in some embodiments be defined asa trajectory for the drone 212 that may be interpreted by control means comprised with thedrone 212 for controlling e. g. electrical motors comprised with the drone 212.

The implementation of the machine learning component 400, as part of e. g. theprocessing circuitry 102 and with ﬁarther reference to Fig. 4, may on a high level bedescribed as an encoder block 402 and a decoder block 404 that may define the backbone ofthe machine leaming component 400. During training, the machine leaming component 400is also dependent on additional software components that may be used to administer data,guide the training process and optimize the machine leaming component 400.

At a greater detail, the above-mentioned software components may comprise: (i) A first module adapted to generate training sequences and labels, batch andat random distribute a sample from the training pool to the machine leaming component toprocess. The labels are sequences of data in same output domain that the machine leamingcomponent 400 will operate in and used as ground truth to obj ectively measure how close orfar away the current iteration of the machine leaming component 400 is to a desired state. (ii) A second module that objectively measure the state of the currentprediction from the machine leaming component 400 and the label that is distributed from thedata administering component defined in (i.) and propagate the individual gradients for theneuron connections backwards to the input neurons. (iii) A third module that will collect the gradients from all connectionsbetween the neurons in the machine leaming component 400 and calculate the individualamount of adjustment needed to better represent the labeled data.

In relation to the machine leaming component 400, the encoder block 402 ispreferably implemented to have an equal amount input neurons as needed input streams ofdata, in other words, the input dimension shall preferably match the input dimension of themachine leaming component 400. The function of the encoder block 402 is here to encodetemporal information from the past and combine with a sample from the input domain, thatis, the output is defined as h(n + 1) = i(n) + h(n - 1) +h(n -2) + + h(n - M) were M is thedefined sequence length of past history needed to step one increment into the future h(n + 1) .

In relation to the above, h(n) represents the output from the encoder block 402 and i(n) 11 represents the input to the encoder block 402. The encoder block 402 preferably consists of Knumber of stacked layers.

The output dimension of the encoder block 402 is deterrnined by the numberof neurons or 'hidden units' in the last of the stacked layers in the encoder block 402.Inforrnation contained in the output from the stacked encoder block 402 is amultidimensional representation of past and current state. The output from the encoder blockis passed on to the decoder block 404, Where the decoder block in tum Will transforrn thelatent representation into tWo parts, firstly generate new input to the encoder block 402,secondly generate an element for the predicted output sequence.

The output sequence may then be generated via sampling from a gaussianmixture model, this due to the multimodal nature of the problem domain. A positive sideeffect from this is that it is also possible to gauge the uncertainty in the prediction yieldinglarger confidence in the predicted data.

In summary, the present disclosure relates to a computer implemented methodfor using machine leaming to predict a future movement associated With a first movableobject, the method comprising the steps of receiving a sequence of representations indicativeof present behavior of the first movable object, Wherein the sequence of representationscomprises a plurality of different behavioral types, forrning a plurality of behavioral maps forthe first movable object by transforrning the plurality of representations for each of theplurality of different behavioral types using at least one machine leaming component, andpredicting the future movement of the first movable object based on the plurality ofbehavioral maps.

By means of the present disclosure, the overall behavior of the first object maybe observed and analyzed for predicting the future movement of the first object.

The control functionality of the present disclo sure may be implemented usingexisting computer processors, or by a special purpose computer processor for an appropriatesystem, incorporated for this or another purpose, or by a hardWire system. EmbodimentsWithin the scope of the present disclo sure include program products comprising machine-readable medium for carrying or having machine-executable instructions or data structuresstored thereon. Such machine-readable media can be any available media that can beaccessed by a general purpose or special purpose computer or other machine With aprocessor. By Way of example, such machine-readable media can comprise RAM, ROM,EPROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium Which can be used to carry or store desired 12 program code in the form of machine-executable instructions or data structures and whichcan be accessed by a general purpose or special purpose computer or other machine with aprocessor. When information is transferred or provided over a network or anothercommunications connection (either hardwired, wireless, or a combination of hardwired orwireless) to a machine, the machine properly views the connection as a machine-readablemedium. Thus, any such connection is properly terrned a machine-readable medium.Combinations of the above are also included within the scope of machine-readable media.Machine-executable instructions include, for example, instructions and data which cause ageneral-purpose computer, special purpose computer, or special purpose processing machinesto perform a certain function or group of functions.

Although the figures may show a sequence the order of the steps may differfrom what is depicted. Also two or more steps may be performed concurrently or with partialconcurrence. Such variation will depend on the software and hardware systems chosen andon designer choice. All such variations are within the scope of the disclo sure. Likewise,software implementations could be accomplished with standard programming techniqueswith rule-based logic and other lo gic to accomplish the various connection steps, processingsteps, comparison steps and decision steps. Additionally, even though the present disclosurehas been described with reference to specific exemplifying embodiments thereof, manydifferent alterations, modifications and the like will become apparent for those skilled in theart.

In addition, variations to the disclosed embodiments can be understood andeffected by the skilled addressee in practicing the claimed present disclosure, from a study ofthe drawings, the disclosure, and the appended claims. Furthermore, in the claims, the word"comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

Claims

1. 1. A computer implemented method for using machine leaming to predict afuture movement associated With a first movable object, the method comprising the steps of: - receiving, at a processing circuitry, a sequence of representations indicative ofpresent behavior of the first movable object, Wherein the sequence of representationscomprises a plurality of different behavioral types, - forrning, using the processing circuitry, a plurality of behavioral maps for thefirst movable object by transforrning the plurality of representations for each of the pluralityof different behavioral types using at least one machine leaming component, and - predicting, using the processing circuitry, the ﬁJture movement of the first movable object based on the plurality of behavioral maps.

2. The method according to claim 1, further comprising the step of:- filtering at least a portion of the plurality of behavioral maps using a predeterrnined filtering scheme.

3. The method according to claim 2, Wherein the predeterrnined filteringscheme is based on the behavioral type or a sensor used for acquiring the sequence of representations.

4. The method according to any one of the preceding claims, Wherein thebehavioral types are selected from a group comprising object pose, head pose, eye gaze, velocity, position, acceleration or an interaction map for the first movable object.

5. The method according to any one of the preceding claims, Wherein adedicated machine leaming component is selected for each of at least a portion of the behavioral types.

6. The method according to any one of the preceding claims, Wherein the atleast one machine leaming component forms its behavioral map based on a present and a predeterrnined number of previous representations for the first movable object. 14

7. The method according to claim 6, Wherein the predeterrnined number of previous representations for the first movable object is dynamic.

8. The method according to any one of the preceding claims, Wherein the stepsof receiving, perforrning and predicting are preforrned using processing circuitry comprised With a second object, the second object being different from the first object.

9. The method according to claim 8, Wherein the second object is movable andthe method further comprises the step of:- forrning control commands for the second movable object based on the predicted future movement of the first movable object.

10. The method according to claim 9, Wherein the control commands represent a trajectory for the second movable object.

11. The method according to any one of claims 9 and 10, Wherein the controlcommands for the second movable object are formed to reduce an interaction between the first and the second object.

12. The method according to any one of the preceding claims, Wherein the first object is a human.

13. The method according to any one of claims 9 - 12, Wherein the second object is at least one of a craft or a vessel.

14. The method according to any one of claims 9 - 12, Wherein the second object is an unmanned aerial vehicle (UAV) or a vehicle.

15. The method according to any one of the preceding claims, Wherein thesequence of representations is based on data generated by an image sensor, a lidar sensor or a radar sensor.

16. The method according to any one of the preceding claims, Wherein thesequence of representations comprises information relating to at least one static object located in a vicinity of the first movable object.

17. A computer system comprising processing circuitry, the computer systemarranged to predict a future movement associated With a first movable object by adapting theprocessing circuitry to: - receive a sequence of representations indicative of present behavior of the firstmovable object, Wherein the sequence of representations comprises a plurality of differentbehavioral types, - form a plurality of behavioral maps for the first movable object bytransforrning the plurality of representations for each of the plurality of different behavioraltypes using at least one machine leaming component, and - predict the future movement of the first movable object based on the plurality of behavioral maps.

18. The computer system according to claim 17, ﬁarther comprising at least one sensor for collecting the sequence of representations.

19. The computer system according to claim 18, Wherein the at least one sensor comprises an image sensor, a lidar sensor or a radar sensor.

20. A second object, comprising a computer system according to any one of claims 17 -19.

21. The second object according to claim 20, Wherein the first object is a human.

22. The second object according to any one of claims 20 - 21, Wherein the second object is movable.

23. The second object according to claim 22, Wherein the second object is at least one of an unmanned aerial vehicle (UAV) or a vehicle. 16

24. The second object according to any one of c1aims 22 - 23, Wherein theprocessing circuitry is further adapted to:- forrn contro1 commands for the second object based on the predicted future movement of the first movab1e object.

25. The second object according to c1aim 24, Wherein the contro1 commands represent a traj ectory for the second object.

26. The second object according to any one of c1aims 24 and 25, Wherein thecontro1 commands for the second object are formed to reduce an interaction between the first and the second object.

27. A computer program product comprising a non-transitory computerreadab1e medium having stored thereon computer program means for operating a computersystem to predict a future movement associated With a first movab1e object using machine1eaming, the computer system comprising processing circuitry, Wherein the computerprogram product comprises: - code for receiving, at the processing circuitry, a sequence of representationsindicative of present behavior of the first movab1e object, Wherein the sequence ofrepresentations comprises a p1ura1ity of different behaviora1 types, - code for forrning, at the processing circuitry, a p1ura1ity of behaviora1 maps forthe first movab1e object by transforrning the p1ura1ity of representations for each of thep1ura1ity of different behaviora1 types using at least one machine leaming component, and - code for predicting, at the processing circuitry, the future movement of the first movab1e object based on the p1ura1ity of behaviora1 maps.