SYSTEM FOR ESTIMATION OF THE CHANGE OF THE OPERATING PHASES OF A TRAFFIC LIGHT SYSTEM, CORRESPONDING VEHICLE AND PROCEDURE

Description

DESCRIPTION of the industrial invention entitled: “System for estimating the change in the operating phases of a plant traffic light, vehicle and corresponding procedure” by: Stellantis Europe SpA, of Italian nationality, Corso Giovanni Agnelli, 200 – 10135 Turin Designated Inventors: Stefano MANGOSIO; Silvano MARENCO; Jacopo MILONE; Alessandro MANCINI Filed on: May 30, 2024 **** DESCRIPTION TEXT Technical field The description refers to the estimation of the change of the operating phases of a traffic light system. Solutions such as those described here are suitable for use, for for example, to improve the conditions of use of the regulation functions automatic traffic light systems. Description of the related technique The measurement and analysis of parameters relating to vehicular traffic conditions for example, in order to actively control such a traffic flow constitute a sector of innovation in continuous development as well as evidenced, for example, by patent documents identified by the string G01C in the International Patent Classification (International Patent Classification or IPC). For example, document FR 3 101 413 A1 describes a procedure and a device for determining the path of a vehicle from one or more routes between a starting point and an arrival point of the vehicle. The time of travel time for each route is estimated based on information representative of the duration of the red and green phases of traffic light systems located along the route. The vehicle speed is calculated so as to allow the vehicle to pass a maximum of green traffic lights so as to reduce the minimum the duration of red lights along the journey by selecting an itinerary from the set of routes determined on the basis of the travel times thus calculated. Document FR 3 113 974 A1 describes a support procedure driving a motor vehicle equipped with means of communication susceptible to receive information relating to the operating status of a traffic light, as well as of a screen and a computer. The procedure involves receiving a duration corresponding to the time remaining before the traffic light changes state and generate data to display on the screen, being able to display a signal driver warning when duration varies discontinuously. Document DE 102017003346 A1 describes a procedure for providing data of the change of phase of operation of a traffic light with a vehicle in approach. The recorded data is transmitted to a data processing unit external to the vehicle. The change in traffic light phase is evaluated in the unit of data processing so as to provide a traffic light switching profile and transmit information to the vehicle about the next phase change traffic light. Document DE 102011077656 A1 describes a solution in which involves turning off the engine of a vehicle approaching a traffic light calculating an estimated time of arrival of the vehicle at the traffic light or a landmark associated with it. Document DE 102020110875 A1 describes a processor circuit of a motor vehicle capable of receiving data indicative of the operating phase of a traffic light and produce a “voting” function related to the relevant phase traffic light. Document CN 102708679 A describes a process for forecasting of short-duration traffic flows at urban traffic-lighted intersections by predicting the transit waiting time at a given point upstream intersection. Document EP 4 042 107 A1 describes a procedure for predicting a speed profile of a vehicle along a predetermined path as a function of input data such as geocoordinates and various other input data such as information on the vehicle's position on a digital map, data relating to the average traffic flow along the route and speed profiles of connected vehicles. The non-patent technical literature is particularly extensive, as well as witnessed (to limit ourselves to just one example also recalled below) to the work of M. Cantas, et al.: “Green Light Optimized Speed ​​Advisory (GLOSA) with Traffic Preview”, SAE Technical Paper 2022-01-0152, 2022. It must therefore be considered completely known that it is possible to associate a traffic light with “smart” data acquisition circuitry configured to acquire signals relating to entities present in the environment and likely to induce a change of the current traffic light phase of the traffic light. All this with the possibility of identifying the relevant parameters (therefore the acquired signals) in a range which, also for the purposes of the solutions described in the subsequently, it can be considered almost unlimited. A problem that can be found when the operating phase of a system traffic light or semaphore (hereinafter, for brevity, “traffic light phase”) is regulated in automatic way is in the fact that, especially in systems equipped with a high degree of “intelligence”, the operating phases of the traffic light (in practice, (the alternation of the red and green phases) can be modified according to of various factors related to the environment and conditions of use. For example, if there is heavy traffic in a certain direction of march, it is possible that the duration of the green phase of a traffic light is extended to facilitate the flow of traffic through the traffic lights in that direction, correspondingly reducing the duration of the red phase. In this way complementary, the duration of the green phase can be reduced in the presence of reduced traffic and / or depending on the fact that a pedestrian requests to be able to cross. The GLOSA system already mentioned above is able to implement regulation functions of this type in a particularly sophisticated way. Such applications, even if seen in relation to the possible use of Human-Machine Interfaces (HMI) on board vehicles, use of Vehicle-to-Infrastructure (V2I) communications and / or for driving self-driving vehicles, they may suffer from the fact that a possible (expected) traffic light phase change may not be detected promptly. In general, the fact that a possible (expected) phase change traffic light is not detected in a timely manner can be the cause of accidents (for example example because a driver has no way of realizing promptly of the fact that the traffic light he is approaching is passing on red to allow the passage of an emergency vehicle) and / or have "forensic" consequences, either in relation to liability for possible accidents, or for the possible contestation of sanctions, for example a fine applied to a driver who had no material opportunity to realize that the traffic light phase has changed from green to red. Purpose and summary One or more forms of implementation aim to contribute to overcome the drawbacks outlined above. According to the solutions described here, this goal is achieved thanks to a system having the characteristics referred to in the following claims. The solutions described here may also concern a corresponding vehicle. A motor vehicle such as a passenger car (possibly self-driving) can be an example of such a vehicle. The solutions described here may also concern a corresponding process that can be implemented by computer, for example by using the capacity processing of an on-board control unit (ECU) of a motor vehicle. Claims form an integral part of the teachings here administered in relation to the forms of implementation of the solutions described here. Brief description of the attached figures One or more forms of implementation will now be described, purely by way of example. non-limiting example, with reference to the attached drawings, where: Figure 1 illustrates a possible application context of the solutions here described, Figure 2 is a block diagram illustrating an example architecture of a system as described here, Figure 3 is a functional block diagram that illustrates in greater detail detail possible forms of implementation of the architecture presented in Figure 2, And Figure 4 is an illustrative functional block diagram of a system of learning that can be used within a system like here described. Detailed description In the following description, various details specific to this purpose are illustrated to provide an in-depth understanding of various examples of embodiments according to the description. The embodiments can be obtained without one or more of the specific details, or with other processes, components, materials, etc. In other cases, known structures, materials or operations are not illustrated or described in detail so that the various aspects of the forms of implementation are not made unclear. A reference to “an implementation form” or “a solution” in the framework of this description is intended to indicate that a particular configuration, structure or feature described in relation to the embodiment is included in at least one form of implementation or solution. Phrases like “in one form of implementation” or “in a solution” which may be present at various points in the This description does not therefore necessarily refer exactly to to the same form of implementation or solutions. Furthermore, particular conformations, structures or characteristics can be combined in any suitable way in one or more embodiments or solutions. The references used here are provided simply for convenience and therefore they do not define the scope of protection or the extent of the forms of implementation. As mentioned, in the various figures, the same references are used for indicate corresponding parts or elements, without repeating the relevant part for brevity's sake description for each figure. Again, the fact that a certain element is indicated as “connected to” or “coupled with” another element, must be understood both in the sense that between these between two elements no other element is interposed, but in the sense that between between these two elements another element can be interposed. When instead it is said that an element is “directly connected” or “coupled directly” to another element, it is understood that between these two elements there is no without interposing any other element. In Figure 1 the reference V indicates a vehicle (a motor vehicle such as (a car, for example) on board which a system 100 is installed like this as described here. The vehicle may be either a vehicle, such as a passenger car, intended for be driven by a human, or a self-driving vehicle. With reference to such a possible application context, the system 100 It may therefore be able to produce both signals that can influence the Vehicle guidance V at the signal level (e.g. visual, acoustic and / or haptic) addressed to the driver of the vehicle, as well as signals that intervene directly on the vehicle control. In Figure 1 the reference TL indicates a system or installation traffic light (in short a “traffic light”) which is assumed here to have an associated AITL “intelligence”, for example in the form of an artificial intelligence system (AI) able to control the operation of the TL traffic light in order to regulate it the operation (mainly the alternation of the red and red traffic light phases) green) as a function of a wide range of parameters representative of the situation of the surrounding environment. The fact of referring here, as regards the functioning of the TL traffic light, at the red and green traffic light phases, without mentioning the phase of yellow, takes into account the fact that (according to methods that vary from country to country) a yellow phase can be predicted either before each of the red phases and of green, either before only one of these phases. For what concerns us here, the yellow phase it can therefore be considered to be effectively incorporated with the red phase. The above mentioned red and green traffic light phases will be in the following recalled by referring primarily to the direction along which the Vehicle V is heading towards traffic light TL to cross it. As can be deduced, for example, from the overview of the prior art contained in the introductory part of this description, the functions of adaptive control of a TL traffic light (hence the implementation of the functions indicated by the AITL block in Figure 1) have experienced over the last years an ever-increasing extension of the range of parameters susceptible to be taken into account and the degree of sophistication of the processing functions carried out. As regards the possible operating criteria of such a system In the case of traffic light control, it is therefore possible to make useful reference to this known technique. For simplicity and clarity of illustration in the following In this description, reference will be made, by way of example, only to some of these parameters (for example, traffic intensity at the traffic light, possible arrival of an emergency vehicle, request to cross by a pedestrian), without this being understood in a limiting sense of the scope of protection. It can also be assumed that the TL traffic light and / or the control system intelligent associated with it (here indicated with AITL) can be in themselves configured to send information (here schematically indicated with the arrow S in Figure 1) indicative of operating conditions, including possible changes in traffic light phases. As mentioned, the solutions described here address the problem related to the fact that such information may eventually not reach or reach too late the vehicle V, thus not allowing the driver or the guidance system autonomously to intervene in the desired way in the regular the progress of vehicle V towards traffic light TL (even beyond it). Solutions such as those described here mainly concern the possibility of equipping the V vehicle with an “on-board” system 100 capable of process -- autonomously -- information on traffic light phases, being able to, for example, determining (even just at the level of estimated probability) whether and when a variation in the phase of the TL traffic light may occur which may affect the Vehicle driving mode V. The fact that the 100 system is an on-board system capable of operating in autonomous way does not exclude that this system can also make use of information such as the information indicated by S (if and when available): the on-board system 100 mounted on the vehicle V however allows the driver of the vehicle V (including an automatic vehicle guidance system) to intervene in a which is not necessarily dependent on the processing functions performed (in general “on the ground”) from the control system indicated with AITL. In solutions as described here, system 100 is assumed to be able to implement processing functions capable of implementing (in a way known per se) a model that reproduces (i.e. simulates or emulates) the processing functions performed from the control system indicated with AITL to regulate the traffic light phases of the traffic light TL. To this end (as further described below) the system 100 can be equipped with a V2X (i.e. Vehicle-to-Vehicle) communication functionality to-Everything) capable of enabling communication between the vehicle V and entities external factors likely to intervene in influencing the driving of the vehicle V. The term V2X may include functionality of the type referred to as V2V (vehicle-to-vehicle), with V2I (vehicle-to-infrastructure, i.e. vehicle-to-infrastructure) or V2P (vehicle-to-pedestrian, i.e. vehicle- a-pedestrian). As further described below, the 100 system is also equipped with “forward-looking” sensor devices (camera, radar, lidar) capable of identifying objects (such as pedestrians) and, in particular, one or more traffic lights TL towards which vehicle V is heading. It can also be assumed that a 100 system like the one illustrated here is capable of exploiting the processing capacity of a processing unit (such as a electronic control unit (ECU) of the type already installed today on currently produced vehicles and capable of being programmed (in a known in itself) to implement procedures capable of evaluating (even just probability level) if and when the TL traffic light operating phases (phases traffic lights) are destined to change, even suddenly, for example following a change in the duration of the green and red phases. This can happen adaptively based on parameters such as (yes Please note that this list is purely illustrative and not exhaustive): information on current and future traffic light phases (red / green) traffic light (for example, how long the traffic light has been green or red), information on the presence of objects (stationary or moving) on ​​the approach path of vehicle V to the traffic light TL or in the vicinity of the same, information on the presence of pedestrians with the possibility (offered by sensors capable of interacting with pedestrians, including body-worn sensors) to provide the intention of a pedestrian to proceed to a crossing or to request the traffic light crossing, traffic flow information, information on the arrival of "special" vehicles, with status of priority traffic (e.g. emergency vehicles, police vehicles) police, vehicles intended for public transport which enjoy priority, for (e.g. tram). Parameters such as those listed above as an example can be either detected directly (for example by “frontal” detectors such as those mentioned above), whether obtained indirectly, for example through a V2X function. Based on these parameters, an “on-board” system such as system 100 is in able to make an estimate (of a predictive type, for example at the probability level) of the possible change in the traffic light operating phases of the TL traffic light. This estimate is carried out by the system 100 autonomously, therefore without having to necessarily depend on the AITL “intelligence” of the TL traffic light and being able to thus preceding the actual change of traffic light phase commanded by the AITL system. A corresponding warning signal issued before and / or in independent of the actual change of the traffic light phase allows the driver of vehicle V (or to an autonomous driving system of vehicle V) of adapt its behavior to the actual operation of the TL traffic light, for example in the implementation of a GLOSA system logic. System 100 is therefore able to implement an estimation model of the (expected) change of traffic light phase, advantageously including an estimate of the amount and type of change expected (for example: lengthening / shortening of the green or red phase and the extent of this lengthening / shortening). In the functional diagram of Figure 2 the references 101A, 101B and 101C indicate various types of sensors / detectors, such as, for example, a camera, a radar and / or a lidar (acronym for Light Detection and Ranging, a sensor capable of determine the position, speed, type and other static characteristics and dynamics of an object using a laser pulse) that can be installed on board the vehicle V which can be seen, as far as it is concerned here, as possible implementation forms of data sensing circuitry configured to be installed on board a land vehicle V so as to produce (first) signals relating to the movement of the vehicle V with respect to at least one TL traffic light and the current traffic light phase of the TL traffic light itself. References 102A, 102B, 102C and 102D indicate the possibility, for a V2X type communication function 104 included in system 100, to receive, through respective interfaces, signals of various nature such as, always by way of non-binding example: via interface 102A, signals indicating the presence of users “vulnerable” (Vulnerable Road User, VRU) such as pedestrians, bicycles, scooters which could interact with the V-vehicle, via interface 102B, “connected” traffic light signals connected to the TL traffic light for carrying out a “Green Wave” or GLOSA type function; via the 102C interface, signals of various nature coming from vehicles “connected” with the vehicle V; and / or via the 102D interface, signals of various nature coming from a road infrastructure (lane indicators, etc.) “connected” to the vehicle V. As far as we are concerned, the V2X communication function and interfaces 102A, 102B, 102C and 102D can be seen as possible forms of implementation of data acquisition circuitry configured to be installed on board the vehicle V and acquire (second) signals relating to entities present in the environment around the vehicle V and likely to induce a change of the current traffic light phase of the TL traffic light. It will also be appreciated that blocks 102A, 102B, 102C, 102D are essentially identifiers of V2X type communication interfaces included in the 100 system, configured - as already mentioned - specifically to be able to operate autonomously on board the vehicle, without having to depend on necessity from information coming from intelligence installed on the ground as well as exemplified by the AITL block in Figure 1. It is recalled once again that the sensors and interfaces mentioned in The above are mentioned purely by way of example and without any limiting intent. This is in relation to the possibility that the system 100 includes devices sensors and / or communication interfaces in smaller or larger numbers compared to the examples cited or sensor devices / interfaces other than those here mentioned as an example. The number and nature of the sensors / interfaces envisaged will obviously be able to vary depending on specific applications or destinations. For example, the VRU 102A interface can be configured so that to be able to detect even signals indicative of the mere intention of a pedestrian to proceed to a crossing, whether in response to an interaction of the pedestrian with a street-level device (for example a push button crossing booking), both in relation to the possibility offered to the pedestrian (and already contemplated in several applications) to have a device portable, e.g. wearable, capable of detecting intent to proceed at the crossing. Likewise, the number and nature of the sensors / interfaces envisaged It may obviously also vary depending on the class of the vehicle itself: as is already the case today, for example, for on-board navigators, vehicles of highest class will be able to have 100 systems equipped with and / or capable of to process a higher number of sensors and interface signals, while on vehicles of the more economical class, 100 systems of a more advanced type may be fitted “basic”. In the diagram of Figure 2 the reference 106 indicates a device of processing (capable of being implemented at least in part at the level of vehicle control unit or ECU) so that it can process (according to the best criteria (described below) output signals intended to regulate the vehicle's driving. This is done through information or warning signals (visual, acoustic and, possibly of a haptic type) presented to the driver of the vehicle on an interface man-machine interface (HMI) 108, both at the level of automatic intervention on the guidance of the vehicle V through an autonomous driving system, a sector which is now constituent a broad area of ​​innovation and development in the automotive sector) 110. For what concerns us here, the device 106 can be seen as expression of the possible implementation of data processing circuitry configured to be installed on board the vehicle and to be coupled to the data sensing circuitry (sensor devices 101A, 101B, 101C) and to the data acquisition circuitry (interfaces 102A, 102B, 102C, 102D and functionality V2X 104) so ​​that it can produce autonomously on board the V vehicle, based on to the signals provided by the data sensing circuitry and to the signals provided by the data acquisition circuitry, prediction signals of possible changes of the traffic light phase of the TL traffic light with respect to the current phase. As seen, these traffic light phase change signals they can be: interface signals capable of being fed to an interface man-machine 108 to be presented to a driver of vehicle V; and / or control signals that can be used to control a guidance apparatus autonomous 110 of the vehicle V. For example, the processing unit 106 may be configured to implement (also using machine learning functions which will be discussed in the final part of this description) an estimation function or model or prediction capable of identifying - on the basis of the detected and / or collected signals - the expected occurrence of a situation likely to produce a change of the TL traffic light phases. For example, the processing unit 106 may be configured to provide a prediction (possibly at the probability level) of when the traffic light phase will be is subject to change, including an estimate of the possible extent or duration of the variation. Such a processing function can be conducted in real time (real time) based on the data / signals coming from sensors 101A-101C and from the interfaces 102A-102D (via V2X function 104). The estimate can be conducted according to the typical methods of operation of a classifier that outputs (towards interface 108 and / or towards the driving assistance system 110) identification signals of the probability of a change in traffic light phase, of the type of change (from green to red or from red to green - it is recalled that, as far as we are concerned here, the phase of the yellow can be seen as incorporated into the red phase, according to the regulations in force in the various countries) and an estimate of the change in duration. It will be appreciated that such a probability value can itself vary – if of the case by bands or range of values ​​- from 0 (0% - event excluded as virtually impossible) to 1 (100% - event taken for granted). These signals can be used at the 108 interface and system levels. of driving 110 in various known ways, for example by emitting warning signals or alarm (of various nature: optical, acoustic or haptic signals) or for intervene directly on vehicle driving at system level 110. Whatever the specific implementation modalities, Figure 3 highlights evidence as the (first) signals provided by the data sensing circuitry 101A, 101B, 101C and the (second) signals provided by the data acquisition circuitry 102A, 102B, 102C, 102D, 104 can be integrated (“fused”) with each other. Such data / signals may be used only as an example. (operating according to criteria which are known to expert technicians in the sector) to develop traffic flow level indications. In this context, operating according to criteria which are known in themselves, it is for example it is possible to identify the development of a situation likely to produce a change of traffic light phases TL. This may occur in response to events such as, for example: the presence, near the TL traffic light, of a pedestrian who intends to cross the intersection and manifest this intent, for example, by pressing a crossing reservation button or with a reaction emotional detectable via a worn sensor, and / or the presence of a bicycle or scooter near the TL traffic light with the intention of crossing the intersection, and / or the arrival of an emergency vehicle, a rescue vehicle, a public transport vehicle with priority, etc., on the route on which it is installed the TL traffic light, and / or the presence of a high level of traffic, such as to suggest, for example increasing the duration of the green phases to encourage flow and clearing a traffic jam. Figure 3 provides a more detailed representation of a possible implementation of the architecture illustrated in Figure 2. Figure 3 highlights some advantageous implementation options of one or more of the blocks represented in Figure 2, also highlighting the fact that the taxonomy exemplified in Figure 2 (with the distinction between circuitry of data collection and data acquisition circuitry) has a functional character and not necessarily structural. In other words, in relation to this second point it is possible to foresee that, for example, the same sensor device 101A, 101B, 101C can be directly feed its detection signal to the processing unit 106, whether cooperate with one or more of the interfaces 102A, 102B, 102C, 102D that acquire (through the V2X function exemplified by block 104) corresponding signals intended to be “fused” with signals coming from the sensors 101A, 101B, 101C with the fusion results fed to the processing unit 106. This fact is exemplified in the diagram of Figure 3 by predicting that the same sensor (for example a 101A camera) and / or the same interface (for example the 102C interface dedicated to connected vehicles) are represented several times, precisely to highlight that the same signal can be used for different purposes. For example, as exemplified in Figure 3, the signal of the camera 101A can be fed to a block 201 together with a signal coming from interface 102B and representing the status of one or more traffic lights connected in such a way as to obtain a merging traffic light data. In this way, in addition to being able to provide, for example, information on objects (fixed or moving, including “vulnerable” subjects) the 101A camera is able to cooperate in provide phase, time and position information relating not only to the first traffic light TL towards which vehicle V is heading, but also in relation to traffic lights subsequent ones being able, for example, to provide a list of traffic lights relating to their position and their present and future phases of operation. For example, a block or module designated 202 can select a particular TL traffic light of interest for vehicle V. Similarly, the signal from the front camera 101A can be used in conjunction with a signal from a 101B forward radar and / or a signal coming from a 101C front lidar and signals of various nature coming from connected vehicles (via the 102C interface of the V2X function) for generate, for example via a 203 merge block, a list of objects (possibly including vulnerable subjects or VRU) who find themselves on the path of the vehicle V with indications relating to their relative position, speed, acceleration, type of objects and, for example, using information coming from connected vehicles, information relating to such vehicles such as information relating to position, speed, type, state (still or moving) with the possible possibility - given by the V2X function - of receiving from such other vehicles “connected” information obtained through the detection functions of which have such vehicles. This with the possibility (for example via an indicated block / form with 204) to carry out - possibly based on selection data coming from from block 202 described above - traffic level information (estimated via a form 205), or the possible presence of vehicles “special” vehicles such as emergency vehicles, police vehicles or transport vehicles public transport that has priority (for example trams on rails). The data / signals from camera 101A, radar 101B, lidar 101C can then be used, depending on signals obtained through interface 102C, to perform a detection action of the type described with reference to connected vehicles with particular attention to connected vehicles likely to be classified as vulnerable subjects (for example by using information from the helmet of a cyclist or scooter rider) thus implementing, in a block indicated with 207, a specific fusion function for vulnerable individuals who are able to provide relevant information once again to the position, velocity, acceleration and other motion data of subjects vulnerable. It will be appreciated that in the case of these subjects it takes on particular interest the ability to detect (for example via body-worn sensors such as electroencephalogram or photoplethysmographic sensors - PPG) character data “emotional” identifiers of intentions or reactions of subjects using such vehicles. As exemplified in Figure 3, module 207 can advantageously also benefit from data coming from the infrastructure connected through the interface labeled 102D. Even in the case of VRU subjects, in a 208 form it is possible to implement an action of selection of VRU subjects based on factors such as the speed, acceleration, type of subject / vehicle identified with the possibility of provide processing unit 106 not only with data relating to the presence of such subjects, as indicated by line 208A, but also indications of a more specific nature emotional detected in a block 209 and transmitted to the unit 106 on a line indicated with 209A. In essence, the representation in Figure 3 exemplifies the possibility, offered in a 100 system thanks to the presence of sensors such as the 101A sensors- 101C and V2X interfaces 102A-102D, to perform (via the blocks indicated with references 201 to 209) data / information fusion processing functions of entry and corresponding selection of objects (including stationary and in vehicles movement and, above all, vehicles that can be classified as vulnerable subjects or VRU) according to various possibilities. For example, at the data fusion level it is possible to perform the following functions: traffic light data fusion, possibly merging information relating to the nearest TL traffic light and information about further traffic lights around the vehicle hosting the 100 system, fusion of information about detected objects with a list of objects surrounding the V-vehicle equipped with the 100 system, and fusion of information relating to entities definable as VRU, yet another time getting a list of such entities surrounding the V-equipped vehicle with system 100, possibly including intentional / emotional data of subjects driving vehicles of this nature. As regards the corresponding selection operation it is possible carry out selection or choice operations such as: the choice or selection of one or more traffic lights of interest that you want to keep track of account in driving the vehicle V, the choice of objects in the environment surrounding the vehicle to be considered interest in driving the vehicle, an estimate of the traffic level not only on the route traveled by the vehicle, but also – and above all – on the roads that cross it, selection of subjects that can be qualified as VRU so as to be able to report both the nature, that the state (still or moving) of such subjects to which it is important to be able to pay particular attention to regulating the movement of the vehicle V, and extension of the assessment relating to subjects such as non-VRU subjects not only to facts, but also to emotional factors such as, for example, intention to cross, the intention to change trajectory. In Figure 3 corresponding parts or elements and parts or elements already described with reference to Figure 2 they are indicated with the same references, without repeating, for simplicity, the relative explanation. It will be appreciated that the processing functions exemplified in Figure 3 by blocks 201 to 209 (such as the various mergers referred to in previously) can be seen as included, together with the "control unit" 106, in the data processing circuitry configured to be installed on board of the vehicle V and to cooperate with the data sensing circuitry (e.g. 101A, 101B, 101C) and with the data acquisition circuitry (e.g. 102A, 102B, 102C, 102D, 104) so ​​as to produce on board the vehicle V, based on the signals of the data sensing circuitry and the signals of the data acquisition circuitry signals predicting the change of the current traffic light phase of at least a traffic light. Also in Figure 3: reference 108 indicates a human-machine interface intended, in function of the signals coming from the circuitry 106, to provide the driver - at least at the probability level - information on the operating phase (red or green) of the TL traffic light of interest both with regards to the current phase and as regards an estimated next phase; reference 110 indicates an autonomous driving system capable of exploit the output information from circuitry 106 to intervene directly in the conduct of the vehicle V. Figure 3 highlights that the data / signals output from the processing unit 106 can advantageously express (at least) three types of information: information on the probability of a traffic light phase change (also short-term); estimate of the remaining duration of the current traffic light phase (it will be appreciated in fact that the change in traffic light phase can eventually translate into even just in the prolongation or shortening of a phase already underway, more that in a phase change), and information on the estimated next phase, to be read in conjunction with as said above: as said, the change of phase does not require a change in the nature of the phase is needed, that is, a transition from red to green or from green to red, but it can also simply manifest itself in a variation in the duration of a phase already in progress, for example in an extension of an ongoing green phase, in order to facilitate the resolution of a traffic jam. It will be appreciated that the solutions described here do not concern, at least in a main, on the methods used in the “intelligence” indicated with AITL in the Figure 1 to determine possible variations in the operating phases of the TL traffic light: in this regard, a rather large number of techniques are known in the art solutions and corresponding procedures, for example “Onda Verde” functionality or Green Light Optimal Speed ​​Advisory (GLOSA) to achieve such functions automatic regulation of traffic light phases. For simplicity of explanation, it can be assumed that system 100 can be equipped, at the level of the modules 201 to 209 and the processing unit 106, with a similar, if not completely comparable, “intelligence” in able to imitate to some extent the processing model that AITL intelligence implements in regulating the traffic light phases of the TL traffic light. A notable feature of the solutions described here is the fact that the system 100 is capable of performing such respective processing functions in a manner autonomous, possibly being able to exchange information with the TL traffic light and the corresponding AITL intelligence without however depending absolutely on such information to determine the driving signals / choices on board the vehicle V with the ability to determine such driving signals / choices as well based on data relating to the kinematics of movement of the vehicle V with respect to the traffic light TL. For further explanation, some simple examples can be given thinking of referring to a vehicle V that is moving at a speed of 15 m / s towards a TL traffic light currently located 90 m from vehicle V. Naturally these quantitative values ​​(and those which will be referred to in the (the following) are purely exemplary in nature and have been chosen mainly for ease of explanation. It can also be assumed that the semaphore TL is initially in a phase green expected to end within 20 seconds with a subsequent transition to a red phase: once again it is recalled that for the purposes of this explanation the yellow phase is to be considered integrated into the red phase. In the 20 seconds in question, in the absence of parameter variations, the vehicle V would be able to travel 300 meters, so as to reach and overtake largely the TL traffic light without interventions on the driving modes. Example 1 Let us now assume, as a first example, that – for example in response to the fact that a pedestrian signals his intent to cross the intersection where he is located the traffic light TL in a transverse direction to the trajectory of the vehicle V - The AITL intelligence of the TL traffic light provides that the switching from green to red should no longer occur after 20 seconds, but earlier, for example after 10 seconds. The system 100 on board the vehicle is configured to reproduce the model of processing (to be considered known in itself) at the basis of the AITL intelligence of the traffic light TL. The system 100 is also able to detect, for example thanks to one of the interfaces, such as interface 102A, the pedestrian's intention to cross and is therefore able to estimate (independently with respect to AITL intelligence) the anticipated change of traffic light phase. At the same time, having the kinematic information on the movement of vehicle V with respect to traffic light TL (detected, for example, by sensors such as the 101A camera - capable of detecting colour as well, therefore the current traffic light phase of the traffic light at the moment, radar 101B or lidar 101C), System 100 is able to establish (at least at the probability level) that, continuing to advance at the same speed for 10 seconds, vehicle V will travel 150 meters, reaching and passing the TL traffic light before the switching from green to red occurs, anticipated in response to the pedestrian's request to cross. Example 2 Now assume that, for example following the detection of the arrival of an emergency vehicle intended to cross the path of the vehicle V at traffic light TL, the change of phase the transition from green to red is anticipated by the intelligence of the AITL traffic light so that it is intended to happen within 6 seconds. By having, for example, via interfaces such as the 102C interface, the related information, the system 100 can estimate (autonomously, being able emulating on board the vehicle V the AITL intelligence of the TL traffic light, but without having to depend on the intelligence of the traffic light or signals from the same) that the transition from red to green is expected to occur within 6 seconds, that is, exactly when the vehicle - which is assumed to continue proceed at a speed of 15 m / s - it will be exactly in correspondence with the stoplight. Presumably, under these conditions vehicle V will be able to continue proceeding, but the unit 106 of the system 100 will be able to send, for example towards the interface man-machine 108, a signal that warns the driver of the vehicle V of the fact which may be found at the TL traffic light just when this will change phase. This will allow the driver of vehicle V to change his driving behavior, for example braking - in safe conditions - in order to avoid finding yourself at the intersection close to the arrival of the rescue vehicle. Example 3 Now suppose that, under conditions similar to those described above, the AITL intelligence of the TL traffic light decides to further anticipate the passage from green to red, expecting this to happen within 5 seconds, i.e. when the vehicle V, assumed to always proceed at the same speed, will find itself at 15 meters from the TL traffic light. Here too the estimate of the (future) phase change traffic light system developed autonomously by the 100 system will be able to allow the system 100 itself, thanks to the fusion of the estimate of the expected moment of traffic light change from green to red and vehicle speed, to warn the driver of the vehicle has the opportunity to brake to avoid engaging the traffic light turned red. Example 4 What was said above is even more true in the case of where the transition from green to red of the TL traffic light (decided by the AITL intelligence and estimated autonomously by the system 100 on board the vehicle V) is intended to take place within 4 seconds, i.e. when vehicle V will be probably still 30 m from the TL traffic light. In this case the system will be able to warn the driver, via the 108 interface, of the importance (in practice of the obligation) of braking in order to avoid engaging the intersection by going through a red light (thus exposing yourself to being fined) and running the risk of colliding with an emergency vehicle, for example. Of course, those given above are just examples. deliberately simplified to illustrate criteria that can regulate the possible operation of the solutions described here. In this regard, it can also be noted that the signals provided by sensors such as the camera 101A, radar 101B and / or lidar 101C allow for the creation of a Evaluation of the kinematics of vehicle V's motion with respect to the traffic light TL more sophisticated than the elementary evaluation presented here as an example being able - possibly in conjunction with other sensors installed on board the vehicle V - obtain more detailed kinematic information regarding position and instantaneous speed of the vehicle and can, for example, also detect data acceleration (the vehicle is accelerating or braking), jerk (for example depending on the driving on an uneven surface and / or the presence of deterrents, etc.). In this way it is possible to make the fusion action of the relevant data with the other data used for processing in unit 106. As regards the methods and characteristics of the messages intended to the driver on the interface 108, a system 100 as described here allows take into account as many possible uncertainties in determining the moment of expected change of the traffic light phase of the traffic light TL (moment - determined - by the AITL intelligence of the TL traffic light and - estimated - by the homologous intelligence implemented in the 100 system on board the vehicle V), how much the possible uncertainty in the determination of one or more parameters is at the level of kinematic parameters, as well as at the level of traffic parameters (parameters environmental, connected vehicles, vulnerable users, etc.). In particular, the signals presented on the interface 108 can take on the character of graphic and / or sound indications of (bands of) probability of the determining a certain event (improbable-low probability-medium event) probable-highly probable-almost certain) so as to allow the driver to adjust his behavior accordingly. This also applies to a possible autonomous driving function as well. as exemplified by block 110, which can be used at the fuzzy logic level probability indications as outlined above, also taking into account the fact that the reaction times of a self-driving vehicle system can be (much) shorter than the reaction times of a human being to vehicle driving V. Figure 4 highlights the possibility of assigning 100 to an on-board system, as described here, (self-)learning characteristics. To illustrate these characteristics, in the left part of Figure 4 the various Blocks 101A-101C, 102A-102D, 104, 201, 203 and 207 are represented simply with a single block that must be understood as grouping itself: data sensing circuitry configured to be installed on board the vehicle V and produce (first) signals relating to the movement of vehicle V with respect to the traffic light TL and the current traffic light phase (red / green) of the traffic light TL, data acquisition circuitry configured to be installed on board the vehicle V and acquire second signals relating to entities present in the environment around the vehicle V likely to induce a change in the traffic light phase current of at least one TL traffic light. Once again it is recalled that the listing and specification of types of sensors / interfaces provided above is to be considered purely exemplary as (also taking into account the class of the vehicle and the specific applications envisaged), the system 100 may comprise a number greater or less than sensors / interfaces or sensors / interfaces of different types from those referred to here by way of example. In this context it will be particularly appreciated that the 201 merger module is configured to receive both information on the current state of the TL traffic light (for example via a front camera 101A) as well as indications on phases, times of operation, positions of one or more traffic lights connected in such a way as to allow to module 202 to generate and send to the processing unit 106 signals indicative of both the state (phase and time) of a first traffic light intended to be encountered by the vehicle, as well as any further traffic lights towards which the vehicle V will be able to go there later. Module 204 can optionally be configured to receive from module 202 indications of the position of the traffic light(s) in question, so as to be able to make a corresponding selection of objects (with indication of position, speed, acceleration and type of such objects) towards module 205 entrusted to the traffic level estimation, while module 206 can, as already mentioned, evaluate similar data, position, speed, type, of any special vehicles present. This can be done starting from position, velocity, acceleration, jerk data of the vehicle and from “environmental” data including emotional data: as seen in the case of entities that can be classified as VRU, the 208 form is able to send to the module 209 corresponding signals that enable module 209 to emit indicative signals even just of behavioral intentions (therefore of behaviors not yet implemented) of VRU entity. Compared to the diagram in Figure 3, the diagram in Figure 4 is illustrative of the possibility of deriving, from the set of sensors / interfaces and modules collectively with the left rectangle in Figure 4, also “tagged” data relating to changes in the duration of traffic light and switching phases such traffic light phases detected "on the field", which can be used to evaluate (both at the level of initial phase of instruction / learning of the system 100, both at the periodical level refinement of the system's learning) the “goodness” of the information provided from processing unit 106 on the basis of data actually found "in the field". This is true both at the initial “offline” learning level, and as regards concerns a possible subsequent "on-the-fly" refinement of the capabilities of assessment (prediction) of the situation by system 100. This can be applied in practice to all parameters sent to the interface. 108 and / or towards a possible autonomous driving system 110, i.e. in the case of the example mentioned above: for the probability of a phase change (going from red to green or from green to red), for the estimation of any variations in the durations of these phases, and / or for the estimation (identification) of the next phase. To carry out these learning phases it is possible to use various Machine Learning (ML) techniques such as, for example: regression trees, support vector machines (SVM), techniques based on long- / short-term memories (Long-Short Term) Memory - LSTM), or various learning techniques currently in use for neural networks (for example with Deep Neural Network techniques). Whatever the specific forms of implementation considered, the self-learning mechanism can be based on comparison operations and comparisons made in modules exemplified here by blocks 301, 302 and 303. Block 301 is an example of a block in which the data (essentially of predictive type) relating to possible changes in the phase of the TL traffic light are compared with the data actually detected by the system itself (which can carry out this operation directly via, for example, the camera 101A) so as to produce a consistency index of the probability of change of the phase as predicted and of the change of phase as actually found “in the field”. Block 302 indicates a block in which a coherence comparison is performed between predicted phases and phases actually encountered (so as to be able to identify, for example, the fact that a predicted change to green was mistakenly indicated as a possible switch to red). Finally the module indicated with 303 is a block configured to create a consistency comparison between the duration of the phases (red and green) as estimated and as actually found in the field. Also in this case the relative information is directly available in the system 100 itself which is able to carry out the corresponding survey "on the field", for example through the front camera 101A. The outputs of blocks 301, 302 and 303 can be powered by a block or module 400 that implements a coherence function generating (according to a metric linked to the procedure implemented in the processing unit 106) a “measure” of the accuracy of the processing function performed by the unit 106. The level of consistency / accuracy determined in block 400 can be submitted in a block 402 (here exemplified as a separate block, but actually susceptible to being incorporated into unit 106) to be compared with an acceptability threshold. When the test of block 402 gives a positive result (Y, indicating a accuracy level compliant with the desired accuracy threshold), the state system current is “frozen” at a step marked 404. This, possibly, in view of a subsequent further verification intended to verify further compliance of the system's functioning to the specifications. Subsequent changes in traffic regulations may in fact, for example, imposing a systematic lengthening of the duration of the phases traffic lights. A negative outcome of the verification step indicated in 402 (N, such as to indicate that the procedure implemented in unit 106 does not give considered results satisfactory and therefore requires to be modified) means that in a subsequent block or module 404 the parameters that regulate the operation of the unit 106 are modified (according to the characteristics of the implemented procedure in unit 106 itself) so as to aim to bring the operation of system 100, once modified, to satisfy, after one or more modifications, the reliability criteria determined in block 400 and verified in block 402. In summary, Figure 4 exemplifies the possibility that the system 100 includes coherence test circuitry (e.g., blocks designated 301, 302, 303, 400, 402) configured to be coupled to the circuitry of data processing (e.g. blocks 201, 202, 204, 205, 206, 207, 208, 209, 106) to perform a comparison of traffic light phase change signals generated by system 100 with signals indicative of the actual change in the traffic light phase of the TL traffic light deduced "on the field". Such coherence test circuits can be configured to take into account I take into account the outcome of the aforementioned comparison so as to: maintain the functioning of the processing circuitry 106 unchanged of system 100, or intervene on the functioning of this processing circuitry 106 for bring the traffic light phase change signals predicted by the system closer to signals indicative of the actual change of the traffic light phase deduced "on the field” from the TL traffic light. Without prejudice to the basic principles, the details of construction and the forms of implementation may vary, even significantly, from what is stated here described and illustrated, purely by way of non-limiting example, without exit the protection area. The scope of protection is determined by the attached claims.

Claims

1. Traffic light phase change estimation system (100) comprising: data sensing circuitry (101A, 101B, 101C) configured to be installed on board a land vehicle (V) to produce first signals relating to the movement of the land vehicle (V) with respect to at least one traffic light (TL) and to the traffic light phase of the at least one traffic light (TL), data acquisition circuitry (102A, 102B, 102C, 102D, 104) configured to be installed on board said land vehicle (V) to acquire second signals relating to entities present in the environment around the vehicle (V) capable of inducing a change in the traffic light phase of the at least one traffic light (TL), and data processing circuitry (201, 202, 204, 205, 206, 207, 208, 209, 106) configured to be installed on board said land vehicle (V), the data processing circuitry (201, 202, 204, 205, 206, 207, 208, 209, 106) being coupled to the data sensing circuitry (101A, 101B,101C) and to the data acquisition circuitry (102A, 102B, 102C, 102D, 104) and configured (202, 205, 206, 208, 209) to produce on board the vehicle (V), based on said first signals and said second signals, signals for predicting the change of the traffic light phase of the at least one traffic light (TL).

2. System (100) according to claim 1, wherein the data sensing circuitry (101A, 101B, 101C) is configured to produce first signals comprising signals selected from signals indicative of position, velocity, acceleration and / or jerk of the land vehicle (V) relative to the at least one traffic light (TL).

3. System (100) according to claim 1 or claim 2, wherein the data sensing circuitry (101A, 101B, 101C) comprises at least one of: a camera (101A), a radar (101B), and / or a lidar (101C).

4. System (100) according to any of claims 1 to 3, wherein the data acquisition circuitry (102A, 102B, 102C, 102D, 104) is configured to acquire second signals relating to entities present in the environment around the vehicle (V) comprising signals selected from signals indicative of: traffic level, presence of vehicles with priority status, and / or presence of vulnerable road users (102A).

5. System (100) according to any of claims 1 to 4, wherein the data acquisition circuitry (102A, 102B, 102C, 102D, 104) is configured to acquire second signals comprising signals selected from: signals from road users (102A) collected via wearable devices; and / or signals from traffic lights located downstream of the at least one traffic light (TL).

6. System (100) according to any of claims 1 to 5, wherein the data acquisition circuitry (102A, 102B, 102C, 102D, 104) includes a V2X communication functionality.

7. System (100) according to any of the preceding claims, wherein the data processing circuitry (201, 202, 204, 205, 206, 207, 208, 209, 106) is configured to produce signals predicting the change of the traffic light phase of the at least one traffic light (TL) selected from: a probability signal indicative of a predicted probability of a change of the traffic light phase of the at least one traffic light (TL); and / or a duration signal indicative of a predicted change of the duration of the traffic light phase of the at least one traffic light (TL); and / or an estimation signal indicative of a new predicted traffic light phase of the at least one traffic light (TL) following the current traffic light phase of the at least one traffic light (TL).

8. System (100) according to any of the preceding claims, wherein the data processing circuitry (201, 202, 204, 205, 206, 207, 208, 209, 106) is configured to produce signals predicting the change of the traffic light phase of the at least one traffic light (TL) comprising signals selected from: interface signals capable of being fed to a human-machine interface (108) to be provided to a driver of said vehicle (V); and / or control signals capable of controlling an autonomous driving apparatus (110) of said vehicle (V).

9. System (100) according to any of the preceding claims, comprising coherence test circuitry (301,302, 303, 400, 402) configured to be coupled to data processing circuitry (201, 202, 204, 205, 206, 207, 208, 209, 106) to perform a comparison (402) of said signals predicting the change of the traffic light phase of the at least one traffic light (TL) with signals indicative of the actual change of the traffic light phase of the at least one traffic light (TL) derived from the at least one traffic light (TL), the coherence test circuitry (301, 302, 303, 400, 402) being configured to, based on the outcome (402) of said comparison: maintain unchanged the operation of said processing circuitry (106),or intervene on the operation of said processing circuitry (106) to bring said signals predicting the change in the traffic light phase of the at least one traffic light (TL) closer to said signals indicating the actual change in the traffic light phase of the at least one traffic light (TL) derived from the at least one traffic light (TL).

10. Land vehicle (V) having installed on board a traffic light phase change estimation system (100) according to any of the preceding claims.

11. Computer-implemented method (106), comprising: - 29 - producing, via detection circuitry (101A, 101B, 101C) on board a land vehicle (V), first signals relating to the movement of the land vehicle (V) with respect to at least one traffic light (TL) and to the traffic light phase of the at least one traffic light (TL), acquiring, via acquisition circuitry (102A, 102B, 102C, 102D, 104) on board said land vehicle (V), second signals relating to entities present in the environment around the vehicle (V) capable of inducing a change in the traffic light phase of the at least one traffic light (TL), and producing, via data processing circuitry (201, 202, 204, 205, 206, 207, 208, 209, 106) a on board said land vehicle (V) coupled to said data sensing circuitry (101A, 101B, 101C) and said data acquisition circuitry (102A, 102B, 102C, 102D, 104),signals for predicting the change of the traffic light phase of the at least one traffic light (TL) based on said first signals and said second signals.