Service station for an intelligent transportation system
The ITS service station aggregates and prioritizes CPMs with improved confidence measures, addressing the processing overload issue in ITS communication architecture, enhancing system efficiency and safety without increasing costs.
Patent Information
- Application Number
- CA3085263
- Authority / Receiving Office
- CA · CA
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2019-07-17
- Filing Date
- 2020-07-02
- Publication Date
- 2026-10-27
- Estimated Expiration
- 2040-07-02
AI Technical Summary
The existing Intelligent Transportation System (ITS) communication architecture faces challenges in managing the overwhelming amount of Collective Perception Messages (CPMs) during heavy traffic situations, leading to processing overload and potential malfunctions or the need for higher processing powers, especially at intersections.
An ITS service station that aggregates CPMs from surrounding ITS-S, calculates a third sensor data value with improved confidence measures, and broadcasts these aggregated CPMs with higher priority, reducing the processing load on receiving ITS-S by prioritizing messages with higher confidence.
This solution significantly reduces processing complexity and load on receiving ITS-S, particularly in heavy traffic scenarios, allowing the use of low-cost systems without compromising safety by leveraging aggregated CPMs with enhanced confidence measures.
Abstract
Description
The present invention relates to a service station for an Intelligent Transportation System (ITS). The basic ITS communication architecture is described in ETSI (European Telecommunications Standards Institute) Standard EN 302 665 and related ETSI standards. A most recent develop- ment in ITS is the so-called Collective Perception Service (CPS) to share information on objects detected by one communi- cation partner, such as a vehicle onboard unit (OBU) or a road- side unit (RSU) (generally called "ITS station", ITS-S), with another communication partner (ITS-S). The CPS in ITS is de- scribed in, e.g., ETSI Technical Report TR 103 562 and ETSI Technical Specification TS 103 324. Figs. 1a - 1c show the present concept of CPS to share "perceptions" (detections, analysis and / or trackings) of ob- jects among participants according to the above-mentioned ETSI standards. In Fig. 1a a vehicle 1 on a road 2 perceives an ob- ject 3, e.g., another vehicle, by means of an own sensor 4 such as a camera, a radar sensor, lidar sensor etc., with a field of view 5. In Fig. 1b the vehicle 1 may be additionally aware of a third vehicle 6 around a bend 7 of the road 2 which obstructs the direct view to the vehicle 6, by means of a wireless commu- nication 8 between an ITS-S aboard the vehicle 6 and an own ITS-S aboard the vehicle 1. In Fig. 1c the third vehicle 6 around the bend 7 perceives a fourth vehicle 9 by means of an own sensor 10 with a field of view 11 and shares information about this perception over the wireless communication 8 with the first vehicle 1. Vehicle 1 thus enjoys the benefit of a "Collective Perception" (CP) from other ITS-S-equipped partici- pants so that it becomes aware of objects beyond its own sensor range even when those objects are not equipped with an ITS-S on their own. The messages exchanged in the CPS to share such percep- tions (here: the message from the ITS-S of the vehicle 10 over the wireless communication 8 to the ITS-S of the vehicle 1) to inform the communication partner (here: the vehicle 1) about the existence, speed, distance, position, direction etc. of a perceived object (here: the vehicle 9) are called Collective Perception Messages (CPMs). Fig. 2 shows the general structure of a CPM as defined in ETSI TR 103 562. The CPM 12 contains - apart from an ITS PDU (Protocol Data Unit) header 13 designat- ing the message as a "CPM package" - a set of CPM parameters 14 in the form of one or more data containers 15 - 18, in particu- lar: - a management container 15 containing basic information about the transmitting ITS-S (in the example of Fig. 1: the ve- hicle 6), such as the station's type (e.g., OBU or RSU) and the station's absolute position; - a station data container 16 with further details, such as the dynamic state or map references of the transmitting ITS- 25 S; - none, one or more sensor information containers 17 de- scribing the sensory capabilities of the transmitting ITS-S, such as range, aperture and position of the sensor / s; and - none, one or more perceived object containers 18 con- taining detailed information about the perceived objects (here: the vehicle 9). Each perceived object is assigned a local ob- ject identifier by the ITS-S generating the CPM which remains constant as long as the object is detected and tracked by the ITS-S. In particular, according to ETSI TR 103 562 a perceived object container 18 may contain sensor data such as distance, speed, acceleration, heading (angle) of a perceived object 9 as measured by the ITS-S's sensor 10, and an indication of the time of measurement of the sensor data. For some data elements, e.g., for distance, speed, angle and object dimension values, ETSI TR 103 562 also provides for confidence measures of the respective data values. The receiving ITS-S can then assess the trustworthiness of the collectively shared perception infor- mation. In general, it is up to the receiving ITS-S to make good use of the wealth of collectively shared sensor data to appro- priately execute road safety applications, such as driver warn- ings or automatic braking and steering functions. However, the wealth of information can overload the processing capabilities of receiving ITS-S in heavy traffic situations, leading either to malfunctions or the need for higher processing powers with increased costs. It is an object of the invention to overcome the shortcom- ings of the prior art and to provide novel devices for improv- ing CPS in ITS. To this end, the invention creates a novel ITS service station, comprising: a receiver having an area of radio coverage and being con- figured to receive a first CPM from a first ITS-S at a first position within the coverage area, the first CPM including first sensor data on an object perceived by the first ITS-S; an aggregator connected to the receiver and configured to aggregate said first sensor data with at least one second sen- sor data on the same object into a third sensor data, which second sensor data either is received via the receiver in a second CPM from a second ITS-S at a second position within the coverage area perceiving the same object or is determined by a sensor of the ITS service station perceiving the same object; and a transmitter connected to the aggregator and configured to broadcast said third sensor data in a third CPM. The novel ITS service station of the invention aggregates CPMs from surrounding ITS-S into aggregated ("third") CPMs so that other ITS-S listening to these broadcasts are eased from the burden of following a multitude of ITS-S and processing a multitude of CPMs. The ITS service station of the invention therefore contributes to reduce the complexity of the CPS for listening ITS-S, in particular when the aggregated CPMs of the ITS service station are prioritized over "normal" CPMs during communication or receipt. A preferred embodiment of the invention is characterized in that the first sensor data includes a first data value and a first confidence measure of said first data value and the at least one second sensor data includes a second data value and a second confidence measure of said second data value, wherein the aggregator is configured to calculate a third data value from said first and at least one second data values and a third confidence measure from said first and at least one second con- fidence measures and to include said third data value and third confidence measure in the third sensor data. An aggregated ("third") data value in the aggregated ("third") CPM will most likely have a better aggregated ("third") confidence measure since it had been aggregated from multiple data sources. Therefore, any ITS-S listening to both "normal" CPMs (here: the first and second CPMs) and "aggregat- ed" CPMs (here: the "third" CPM of the inventive ITS service station) can choose to process and consider the CPM showing the best confidence measure for a specific data value needed, lead- ing to an implicit prioritizing of the CPMs of the ITS service station at the receiving ITS-S. The receiving ITS-S may ignore sensor data regarding the same object from all other CPMs in favor of the sensor data on this object in the aggregated CPM. Processing load in the receiving ITS-S is thus significantly reduced, in particular in heavy traffic situations, e.g., at an intersection, and low-cost ITS-S with modest processing capa- bilities can be used without compromising safety. In one embodiment, the aggregator is configured to deter- mine objects in relation to sensor data to be the same when the objects match in one or more of object positions, speeds, head- ings, and accelerations, as indicated in the respective CPMs. This leads to accurate matching results, however, requires that the respective object positions are either geo-referenced to a common or global coordinate system, e.g., given as absolute geo-coordinates, or map-matched to streets, places, landmarks etc. in a map. Alternatively - or additionally, for added accuracy and reliability - the aggregator is configured to determine objects in relation to sensor data to be the same when the objects match at least in object appearance characteristics indicated in the respective CPMs. Appearance characteristics can be any one or more of: an object dimension, colour, shape, orientation etc. The mapping as to object appearance can be useful when ob- ject positions are indicated relatively to a perceiving ITS-S or are not accurate enough to distinguish between objects close to each other. Appearance and position matching can also be combined to improve the accuracy of the match. The third sensor data may additionally include the number (count) of first and second sensor data from which the third sensor data has been aggregated, and / or may additionally in- clude the number (count) of first and second ITS-S from whose CPMs the third sensor data has been aggregated. An ITS-S re- ceiving the aggregated CPM can use this information to further assess the confidence of a sensor data value indicated therein. According to a further preferred feature of the invention each sensor data includes a local identifier of the object re- lated to said sensor data and the aggregator is configured to assign a global identifier to all local identifiers relating to the same object and to include that global identifier in the third CPM. Receiving ITS-S can then use the global object iden- tifiers in, e.g., own CPMs sent to other participants. If the assignment table between local and global identifiers is dis- seminated from the ITS service station to the receiving ITS-S, too, e.g., within the aggregated CPM or in a separate broadcast message, then receiving ITS-S may more easily match the sensor data on an object from the aggregated CPM with the sensor data on the same object from other "normal" CPMs without the need of own position and / or appearance matches to determine the identi- ty of objects over different CPMs. For keeping inventory and tracking of objects in its area of coverage over time, the aggregator of the ITS service sta- tion may have a memory for storing first and second CPMs in- cluding timestamps of the sensor data therein and may be con- figured to retrieve, for aggregating said third CPM, all sensor data from the memory having timestamps falling within a select- ed period of time. Although the aggregated CPMs of the ITS service station of the invention may implicitly have priority over "normal" CPMs in that they will usually carry sensor data with higher confi- dence measures than the normal CPMs of other ITS-S, the aggre- gated CPMs of the ITS service station may additionally be flagged with a higher priority than normal CPMs. This may be done by, e.g., including a "high priority" flag in the header of the aggregated CPM. Receiving ITS-S then do not need to com- pare confidence measures to prioritize aggregated CPM over nor- mal CPMs, but just will look for the high priority flag, to speed up processing. The ITS service station of the invention can either be moveable, e.g., in the form of an onboard unit on a vehicle, or stationary, such as a roadside unit or infrastructure. In a particularly preferred embodiment the ITS service station is a roadside unit at an intersection. At intersections high vehicle traffic and hence communication traffic is to be expected so that receiving ITS-S benefit most from the load-reducing and safety-increasing CPM aggregation service of the inventive ITS service station. The invention will now be described in further detail by means of exemplary embodiments thereof under reference to the enclosed drawings, in which show: Figs. 1a - 1c the basic concept of Collective Perception in ITS according to the state of the art in three plan views of an exemplary road section; Fig. 2 a data structure of a CPM according to the state of the art; Fig. 3 an ITS service station according to the invention at an intersection in a plan view; Fig. 4 an ITS service station according to the invention in a block diagram; Fig. 5a an exemplary memory contents of the ITS service station of Fig. 4; Fig. 5b an example of an aggregated CPM as broadcast by the ITS service station of Fig. 4; and Fig. 5c an exemplary object identifier assignment table in the memory of the ITS service station of Fig. 4. Figs. 1 and 2 referring to the CPS in ITS and the CPM data structure, respectively, have been explained at outset. Fig. 3 shows an ITS service station 20 mounted stationari- ly as a roadside unit (RSU) at one of the corners of an inter- section S between four roads A, B, C, D. Four exemplary vehi- cles 21 - 24 are shown approaching or just about to enter the intersection S. The positions of the vehicles 21 - 24 on the roads A - D are designated as <semantics>PA<annotation encoding="application / x-tex">P_{A}< / annotation>< / semantics>, <semantics>PB<annotation encoding="application / x-tex">P_{B}< / annotation>< / semantics>, <semantics>PC<annotation encoding="application / x-tex">P_{C}< / annotation>< / semantics> and <semantics>PD<annotation encoding="application / x-tex">P_{D}< / annotation>< / semantics>, respectively. The vehicles <semantics>21−23<annotation encoding="application / x-tex">21 - 23< / annotation>< / semantics> each carry an ITS-S <semantics>25−27<annotation encoding="application / x-tex">25 - 27< / annotation>< / semantics> in the form of an onboard unit (OBU). Vehicles 22, 23 are exemplarily equipped with a sensor 28 with a respective field of view 29, capable of perceiving an object (here: the fourth vehicle 24). The sensors 28 may be of any kind, e.g., a camera, a radar or lidar sensor, an acoustic sensor, a vibration sensor, an infra- red sensor etc. The ITS service station 20, too, may have an own sensor 28 to perceive objects 24 in its vicinity, although this is not obligatory. Generally speaking, each of the ITS-S 25 - 27 and ITS service station 20 may have none, one or more sensors 28, also of different sensor types. Instead of being stationarily mounted as a roadside unit, the ITS service station 20 could also be mobile, e.g., aboard a vehicle as an OBU. The object 24 perceived by the sensors 28 may be of any kind, e.g., a manned or unmanned land, sea or air vehicle, a pedestrian, an animal, a machine, a traffic sign, a radio, a light or infrared beacon broadcasting some kind of information which is useful to be collectively perceived, and the like. In the traffic scenario depicted in Fig. 3, the vehicle 21 cannot see the object (here: vehicle) 24 approaching the inter- section S since its view is blocked by buildings 30 at the cor- ner of the roads A and D. However, the vehicles 22, 23 share their perception of the vehicle 24, as detected by their sen- sors 28, via CPMs 31, 32 sent from their ITS-S 26, 27 (the "perceiving" or "disseminating" ITS-S) to the ITS-S 25 (the "receiving" or "listening" ITS-S) of the vehicle 21. These "normal" CPMs 31, 32 are also received by the ITS service sta- tion 20, which creates an "aggregated" CPM 33 therefrom, as follows. With reference to Figs. 3 and 4, the ITS service station 20 has a receiver 34 with an area of radio coverage 35 to re- ceive the CPMs 31, 32 from the ITS-S 26, 27 in its neighbor- hood. It goes without saying that the radio coverage area 35 will be dependent both on the transmitting power of the dissem- inating ITS-S 26, 27 and the receiving sensitivity of the re- ceiver 34. For ease of description, the various normal CPMs 31, 32 ... are designated as <semantics>CP1<annotation encoding="application / x-tex">CP_1< / annotation>< / semantics>, <semantics>CP2<annotation encoding="application / x-tex">CP_2< / annotation>< / semantics>, ..., generally <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics>, in the fol- lowing. An aggregator 36 connected to the receiver 34 processes the set {CPm} of the received CPMs CPm and calculates the ag- gregated CPM 33, called <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics> in the following, therefrom. The aggregated CPM <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics> is then broadcast by a transmitter 37 con- nected to the output of the aggregator 36 so that it can be re- ceived by listening ITS-S in the vicinity, such as (here) the ITS-S 25 on the vehicle 21. The transmitter 37 and the receiver 34 of the ITS service station 20 can be implemented by a com- bined transceiver, too. To calculate the aggregated CPM CP2 from the received nor- mal CPMs CPm the aggregator 36 has a memory 38 in which - among other programs and data as needed - two tables are stored: a table 39 storing the CPMs <semantics>CPmℓ<annotation encoding="application / x-tex">CP_{m_{\ell}}< / annotation>< / semantics> shown in Fig. 5a, and an (op- tional) table 40 storing assignments between global and local object identifiers, shown in Fig. 5c and described further down below. With reference to Figs. 2 and 5a a CPM 12, 31, 32 or <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics>, respectively, contains - apart from the other data depicted in Fig. 2 - in any perceived object container 18, here called <semantics>oc1<annotation encoding="application / x-tex">oc_1< / annotation>< / semantics>, <semantics>oc2<annotation encoding="application / x-tex">oc_2< / annotation>< / semantics>, ..., generally <semantics>ocn<annotation encoding="application / x-tex">oc_n< / annotation>< / semantics>, a local object identifier id, usually assigned by the perceiving ITS-S 26, 27, and one or more sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics> (i = 1, 2, ...) on the object 24 with the local object identifier idn. The sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics> on an object 24 perceived by a dissem- inating ITS-S 26, 27 may contain any data value di derived from an output of one or more sensor / s 28, e.g., a distance of the object 24 to the sensor 28, a speed of the object 24, a geo- referenced or map-matched position PD of the object 24, a head- ing, angle or path of travel of the object 24, one or more di- mensions of the object 24, a shape, color or class of the ob- ject 24 as determined by the sensor 28, e.g., as taken by a camera and determined by image processing, etc. For example, any of the data items in the perceived object container 18 of a CPM according to ETSI TR 103 562 can be the data value d, such as the data elements xDistance, yDistance, zDistance, xSpeed, ySpeed, zSpeed, xAcceleration, yAcceleration, zAcceleration, yawAngle, planarObjectDimension1, planarObjectDimension2, ver- ticalObjectDimension, objectRefPoint, dynamicStatus, Classifi- cation, MatchedPosition according to ETSI TR 103 562. Some of the data values di which are provided by the re- spective sensor 28 or a suitable processor connected to the sensor / s 28 in the ITS-S 26, 27 or the ITS service station 20 may be provided with a confidence measure cfi, in particular the distance, speed, angle and dimension data values di. For such data values di, the respective sensor data sdi is a pair (d, cf); comprised of the data value d; and the associated con- fidence measure cfi. The confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics> of a data value <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics> may be any statistical measure of the confidence, reliability, trustwor- thiness, non-error rate etc. of this data value <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>. For exam- ple, the confidence measure cfi can be the 95%-confidence in- terval of the respective data value di, i.e., that with a prob- ability of 95% the data value falls within this interval. Of course, other measures of confidence could be used as explained later on. If the ITS service station 20 has one or more own sensors 28 which generate their own sensor data <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics> (k = 1, 2, ...), the output of these sensors 28 can, e.g., be stored - in the same format as the received CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> - in data records <semantics>SD1<annotation encoding="application / x-tex">SD_1< / annotation>< / semantics>, <semantics>SD2<annotation encoding="application / x-tex">SD_2< / annotation>< / semantics>, ..., generally <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics>, for example in the same table 39, as shown in Fig. 5a. From at least two received CPMs CPm, or at least one re- ceived CPM CPm and at least one sensor data record SDk, the ag- gregator 36 calculates the aggregated CPM <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics> as follows. As shown in Fig. 5b, an aggregated sensor data <semantics>sdΣ<annotation encoding="application / x-tex">\operatorname{sd}_{\Sigma}< / annotation>< / semantics> can be any aggregation of sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics> of the same sensor type (i.e., the same type of sensor 28 or combination of sensors 28) from at least two different CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> or at least one <semantics>CPM<annotation encoding="application / x-tex">CPM< / annotation>< / semantics> <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> and at least one record <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics>. By means of "aggregation" any mathematical operation or representation which can be performed on two or more data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics> and <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> which yields one data val- ue <semantics>dπ<annotation encoding="application / x-tex">d_{\pi}< / annotation>< / semantics> shall be understood. For example, the aggregated data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> can be an average or weighted average of the individu- al data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> of which it has been aggregated. Alterna- tively, the aggregated data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> could be that one of the originating data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> which has the best confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_k< / annotation>< / semantics>, respectively. Or, the aggregated data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> is an average or weighted average of just the two or three data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> with the best confidence measures. Furthermore, for a weighted average the individual data values <semantics>di,<annotation encoding="application / x-tex">d_{i,}< / annotation>< / semantics> <semantics>dk<annotation encoding="application / x-tex">d_{k}< / annotation>< / semantics> could be weighted by their respective confidence measures <semantics>cfi<annotation encoding="application / x-tex">cf_{i}< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_{k}< / annotation>< / semantics>. As an example, one embodiment of an aggregation function F for a data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> formed from at least two different CPMs <semantics>CPi<annotation encoding="application / x-tex">CP_{i}< / annotation>< / semantics> (i = 1 ... N) could be: [Image disponible dans le document PDF, Image available in the PDF document] (1) Of course, other aggregation functions F could be used such as, for example, using only the data value di with the "best" confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics> as the aggregated data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics>, using step functions or a binary decision tree to select one or more data value / s di with "good" confidence measure / s surpas- sing a specified threshold while discarding other ones below the threshold, including other information such as the number (count) of data sources (stations 20, 26, 27 and / or sensors 28), their positions, speeds and / or headings with respect to the object 24, the field of view 29 of the respective sensors 29, etc. Furthermore, also sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics> from more than one sensor 28 of an ITS-S 26, 27 and / or the ITS service station 20 which relate to the same object 24 can be aggregated into one aggregated sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> for that object 24, e.g., by combin- ing different dimension values into a shape value, or combining speed and heading values into a movement vector, etc. If data values <semantics>di<annotation encoding="application / x-tex">d_{i}< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_{k}< / annotation>< / semantics> with respective confidence measures <semantics>cfi<annotation encoding="application / x-tex">cf_{i}< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_{k}< / annotation>< / semantics> are used in the aggregation function F the data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> could be weighted with their respective confidence measures <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_k< / annotation>< / semantics>, for example as follows: [Image disponible dans le document PDF, Image available in the PDF document] (2) The aggregated sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> can have an aggregated con- fidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> attributed to the aggregated data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics>. Any statistical operation or measure can be applied to calcu- late this aggregated confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics>. The type of calcu- lation also depends on which type of confidence measure is used for the confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_{i}< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_{k}< / annotation>< / semantics> in the CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_{m}< / annotation>< / semantics> and data records SDk. For example, the confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_k< / annotation>< / semantics> could be in- dicated in the form of a confidence interval into which a given percentage of all readings of a sensor 28 falls, e.g., a confi- dence interval for a percentage (confidence level) of 95%. Or, the other way round, the confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_k< / annotation>< / semantics> could be indicated as the percentage (confidence level) of all measure- ments of a sensor 28 which fall into a given confidence inter- val (error range). All sorts of known statistical measures for indicating such confidence measures can be used. For example, if the confidence measure <semantics>cfi<annotation encoding="application / x-tex">cf_i< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_k< / annotation>< / semantics> is indi- cated as that confidence interval into which 95% of all meas- urements fall (the "95%-confidence level"-confidence interval) as used in ETSI TR 103 562 for indicating the confidence of the data values distance, speed, angle and dimension, and if we as- sume the sensors 28 to have a Gaussian error distribution, for the exemplary aggregation function F given in equation (2) the aggregated confidence measure cfE can be calculated using [Image disponible dans le document PDF, Image available in the PDF document] (3) as [Image disponible dans le document PDF, Image available in the PDF document] (4) wherein <semantics>Φ<annotation encoding="application / x-tex">\Phi< / annotation>< / semantics> designates the Cumulative Distribution Function and <semantics>Φ−1<annotation encoding="application / x-tex">\Phi^{-1}< / annotation>< / semantics> designates the inverse thereof. When other aggregation functions F than that of equation (2) are used, the aggregated confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> - be it in- dicated as a confidence level for a given confidence interval or as a confidence interval for a given confidence level - can be calculated accordingly, as known in the art of statistics and error propagation calculus. The aggregated confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> can also be a compo- site field or concatenation of the confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> as calculated above and other information such as the number (count) of data sources responsible for that confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics>, their positions, speeds and / or headings with respect to the object 24, the fields of view 29 of the respective sensors 28, etc. For example, the more different the positions of the data sources, i.e. the positions of the sensors 28 and / or the posi- tions of the stations 20, 26, 27, with respect to a specific object 24 are, the better the data quality of the data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> aggregated therefrom is. To be able to aggregate the originating data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> or originating sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics> (including the respective confidence measures <semantics>cfi<annotation encoding="application / x-tex">cf_{i}< / annotation>< / semantics>, <semantics>cfk<annotation encoding="application / x-tex">cf_{k}< / annotation>< / semantics>, if applicable) into the aggregat- ed data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> or sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics>, respectively, the aggrega- tor 36 has to determine that the originating data values or sensor data all relate to the same perceived object 24. This is easy if the object identifiers id in the respective CPM <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> and / or data records <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics> are the same. This may happen when the disseminating ITS-S 26, 27 (and the ITS service station 20, if provided with an own sensor 28) already use "global" object identifiers instead of "local" object identifiers. The term "local" object identifier refers to an object identifier id which had been assigned by the respective station 26, 27, 20 itself (locally), whereas the term "global" object identifier, here designated g-id in the following, refers to an object identifier which had been assigned by the ITS service station 20 at least "area wide" in its area of coverage 35. To this end, the ITS service station 20, and in particular the aggregator 36, optionally hosts and manages the assignment table 40 in the memory 38, storing an association (assignment) between a global object identifier g-idn and all local object identifiers <semantics>idm,n<annotation encoding="application / x-tex">id_{m,n}< / annotation>< / semantics> regarding the same perceived object 24, as de- scribed in a perceived object container ocn, from a multitude of different CPMs CPm received (plus, if applicable, of differ- ent data records <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics> stored). Fig. 5c shows an example of the assignment table 40. The aggregator 36 may disseminate the table 40 in its aggregated CPMs CP2 or in separate broadcasts such as CAMs (Common Aware- ness Messages) or BST (Beacon Service Table) messages to all listening ITS-S 25 - 27 in its area of coverage 35. Listening ITS-S 25 - 27 may then use the disseminated global object iden- tifiers <semantics>g−idn<annotation encoding="application / x-tex">g-id_n< / annotation>< / semantics> in their CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> to the ITS service station 20, so that matching the data values <semantics>di<annotation encoding="application / x-tex">d_i< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_k< / annotation>< / semantics> or sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics>, respectively, as to the "same" object 24 can be done by looking for the same global object identifier <semantics>q−idn<annotation encoding="application / x-tex">q-id_n< / annotation>< / semantics>. On the other hand, the aggregator 36 can determine itself whether objects 24 about which sensor data sdi is communicated in the CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> (or sensor data <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics> is recorded in the records <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics>) relate to the same object 24. To this end, sensor data sdi, sdk indicative of object positions - be they given "abso- lutely" in terms of a global or at least geo-referenced coordi- nate system or "relatively" to the position of the respective disseminating ITS-S 26, 27 and then converted into global or geo-referenced coordinates or map-matched position data - may be matched to each other so that congruency (within a certain error margin) in position indicates the same object. Alterna- tively or additionally, further movement characteristics of the object 24 can be taken into account in the match, such object speed, heading, and / or acceleration. Alternatively, object identity could be determined by the aggregator 36 by means of analyzing sensor data sdi, sdk indic- ative of one or more characteristics of an object's appearance, such as one or more dimension / s of an object 24, its color, shape, orientation, heading etc. Both methods of matching, as to object positions and as to object appearance characteristics, can be combined to increase the accuracy of the match. Based on the matches found, the aggregator 36 can then as- sign the global object identifier g-idn to the local object identifiers <semantics>idi<annotation encoding="application / x-tex">id_i< / annotation>< / semantics>, <semantics>idk<annotation encoding="application / x-tex">id_k< / annotation>< / semantics> of the matching objects 24. The aggregation performed by the aggregator 36 may take into account timing aspects. Each CPM <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> and record <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics>, and in particular each perceived object container ocn or even each individual sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics>, may contain a timestamp t in- dicative of the time of measurement of the respective sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_{i}< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_{k}< / annotation>< / semantics>. The timestamp t may be indicated in any suitable format, be it relatively to a time of sending the respective <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> or the time of storing the respective record <semantics>SDk<annotation encoding="application / x-tex">SD_k< / annotation>< / semantics>, or abso- lutely in terms of a systemwide reference clock. The timestamps t can also take into account the track or estimated movement of a perceived object 24, and any calcula- tion, processing or transmission delays. In this way, the "age" of a sensor data <semantics>sdi,j<annotation encoding="application / x-tex">sd_{i,j}< / annotation>< / semantics> <semantics>skk<annotation encoding="application / x-tex">sk_k< / annotation>< / semantics> can be accounted for by the aggrega- tor 36 when matching objects 24 for identity and / or calculating the aggregated CPM <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics>. For example, the aggregator 36 may, when aggregating the CPM <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics>, only use sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_{i}< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_{k}< / annotation>< / semantics> from its memory 38 whose timestamps t fall within a selected period of time, for example into a past cycle interval, when the ITS service station 20 cyclically sends CPMs <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics>. An aggregated sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> may even relate to a "glob- al" object confidence of the perceived object 24 taking into account sensor data <semantics>sdi<annotation encoding="application / x-tex">sd_i< / annotation>< / semantics>, <semantics>sdk<annotation encoding="application / x-tex">sd_k< / annotation>< / semantics> of different (sensor) types. In the aggregated CPM <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics> the aggregator 36 may optionally include the number (count) of originating sensor data <semantics>sdi,j<annotation encoding="application / x-tex">sd_{i, j}< / annotation>< / semantics> <semantics>sdk<annotation encoding="application / x-tex">sd_{k}< / annotation>< / semantics> from which a specific aggregated sensor data sd2 had been ag- gregated, and / or the number (count) of disseminating ITS-S 26, 27 from whose CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> that specific aggregated sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> had been aggregated. The numbers (counts) can be, e.g., at- tached as data field / s to the respective aggregated confidence value / s cf<semantics>Σ<annotation encoding="application / x-tex">_{\Sigma}< / annotation>< / semantics> in the aggregated CPM CP<semantics>Σ<annotation encoding="application / x-tex">_{\Sigma}< / annotation>< / semantics>. These numbers (counts) can then be used by a receiving ITS-S 25 to select if or which one of several received aggregated CPM <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics> is to trust most re- garding a specific sensor data. Usually, the receiving ITS station 25 will select and use that aggregated sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> on an object 24 which has the best confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> attributed to it, e.g., the small- est confidence interval or the highest confidence level, when the confidence measure is expressed in such terms. However, with the additional knowledge of the numbers (counts) of origi- nating sensor data or ITS-S, from which the sensor data sd had been aggregated, the receiving ITS-S 25 can improve the selec- tion, e.g., by weighting the confidence measures <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> by their respective numbers (counts) of underlying data. On the other hand, said numbers (counts) may be particularly useful for sen- sor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> which does not comprise a confidence measure <semantics>cfΣ<annotation encoding="application / x-tex">cf_{\Sigma}< / annotation>< / semantics> at all. For example, it the data value <semantics>dΣ<annotation encoding="application / x-tex">d_{\Sigma}< / annotation>< / semantics> in the sensor data <semantics>sdΣ<annotation encoding="application / x-tex">sd_{\Sigma}< / annotation>< / semantics> is an average of the originating data values <semantics>di<annotation encoding="application / x-tex">d_{i}< / annotation>< / semantics>, <semantics>dk<annotation encoding="application / x-tex">d_{k}< / annotation>< / semantics>, the number (count) of averaged values is a measure of the quality of the process of averaging. The aggregated CPMs <semantics>CPΣ<annotation encoding="application / x-tex">CP_{\Sigma}< / annotation>< / semantics> broadcast by the ITS service sta- tion 20 as CPMs 33 may optionally contain a "high priority" flag or a flag indicating a priority which is higher than those of the "normal" CPMs <semantics>CPm<annotation encoding="application / x-tex">CP_m< / annotation>< / semantics> (CPMs 31, 32). ITS-S 24 listening to the CPMs 31, 32 may prefer CPMs 33 with higher priority (or disregard CPMs 31, 32 with lower priority) so that the aggre- gated CPMs 33 of the ITS service station 20 are favored. The invention is not restricted to the specific embodi- ments disclosed herein, but encompasses all modifications, var- iants and combinations thereof that fall within the scope of the appended claims.
Claims
1. Intelligent Transportation System, ITS, service sta- tion, comprising: a receiver having an area of radio coverage and being con- figured to receive a first Collective Perception Message, CPM, from a first ITS station at a first position within the cover- age area, the first CPM including first sensor data on an ob- ject perceived by the first ITS station; an aggregator connected to the receiver and configured to aggregate said first sensor data with at least one second sen- sor data on the same object into a third sensor data, which second sensor data either is received via the receiver in a second CPM from a second ITS station at a second position within the coverage area perceiving the same object or is de- termined by a sensor of the ITS service station perceiving the same object; and a transmitter connected to the aggregator and configured to broadcast the third sensor data in a third CPM, wherein the first sensor data includes a first data val- ue, the at least one second sensor data includes a second data value and the aggregator is configured to calculate a third data value from said first and at least one second data values and to include said third data value in the third sensor data, wherein each of the first, second and third data values is a - 23 - distance of the object with respect to a respective sensor, a speed of the object, an acceleration of the object, a geo- referenced or map-matched position of the object, a heading, angle or path of travel of the object, one or more dimensions of the object, a shape, color or class of the object.
2. ITS service station according to claim 1, wherein the first sensor data includes a first confidence measure of said first data value and the at least one second sensor data in- cludes a second confidence measure of said second data value, wherein the aggregator is configured to calculate a third con- fidence measure from said first and at least one second confi- dence measures and to include said third confidence measure in the third sensor data.
3. ITS service station according to claim 1 or 2, char- acterized in that the aggregator is configured to determine objects in relation to sensor data to be the same when the ob- jects match in one or more of object positions, speeds, head- ings, and accelerations.
4. ITS service station according to any one of claims 1 to 3, characterized in that the aggregator is configured to determine objects in relation to sensor data to be the same when the objects match at least in object appearance charac- teristics.
5. ITS service station according to any one of claims 1 to 4, characterized in that the third sensor data includes a count of first and second sensor data from which the third sensor data has been aggregated.
6. ITS service station according to any one of claims 1 to 5, characterized in that the third sensor data includes a count of first and second ITS stations from whose CPMs the third sensor data has been aggregated.
7. ITS service station according to any one of claims 1 to 6, characterized in that each sensor data includes a local identifier of the object related to said sensor data, wherein the aggregator is configured to assign a global identifier to all local identifiers relating to the same object and to in- clude that global identifier in the third CPM.
8. ITS service station according to claim 7, character- ized in that the aggregator is configured to store the assign- ment between the global identifier and the local identifiers in an assignment table and to broadcast said assignment table in the third CPM or in a separate message.
9. ITS service station according to any one of claims 1 to 8, characterized in that the aggregator has a memory for storing first and second CPMs including timestamps of the sen- sor data therein and is configured to retrieve, for aggregat- ing said third CPM, all sensor data from the memory having timestamps falling within a selected period of time.
10. ITS service station according to any one of claims 1 to 9, characterized in that the third CPM is flagged with a higher priority than the first and second CPMs.
11. ITS service station according to any one of claims 1 to 10, characterized in that the ITS service station is a roadside unit.
12. ITS service station according to claim 11, character- ized in that the ITS service station is a roadside unit at an intersection.