AUTONOMOUS DRIVING BASED ON DISTANCE AND SPEED OF CERTAIN DESTINATIONS
Patent Information
- Application Number
- DE602020041602
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2019-02-12
- Filing Date
- 2020-01-24
- Publication Date
- 2024-11-20
- Estimated Expiration
- 2040-01-24
AI Technical Summary
Current autonomous driving methods face safety deficiencies when multiple target objects are detected, as they often rely on the closest object for adaptive speed regulation, which can lead to risky maneuvers if that object behaves dangerously, such as quickly overtaking a slower vehicle.
A method for autonomous driving that regulates the ego-vehicle's speed based on the second target object while maintaining a target distance from the first target object, reducing the risk of sudden maneuvers and ensuring passenger safety by avoiding dangerous proximity to the first target object.
This approach significantly reduces the risks associated with multiple vehicles traveling at different speeds, maintaining a safe distance and avoiding sudden avoidance maneuvers, thereby enhancing passenger comfort and safety.
Description
[0001] The present invention belongs to the field of autonomous vehicles. It relates in particular to a method and a device for autonomous driving of a vehicle, called an ego-vehicle, for managing the situation where two target objects concern the ego-vehicle.
[0002] It is particularly advantageous in the case of a motor vehicle whose speed is adaptively regulated by following target objects.
[0003] A "vehicle" means any type of vehicle such as a motor vehicle, a moped, a motorcycle, a storage robot in a warehouse, etc. "Autonomous driving" of an "autonomous vehicle" means any process capable of assisting the driving of the vehicle. The process may thus consist of partially or totally steering the vehicle or providing any type of assistance to a natural person driving the vehicle. The process thus covers all autonomous driving, from level 0 to level 5 in the OICA scale, for International Organization of Motor Vehicle Manufacturers.
[0004] In the following, the term "ego-vehicle" is used to designate the vehicle whose autonomous driving is determined.
[0005] A "road" means any means of communication involving the physical movement of a vehicle. Examples of roads include a national, departmental, local, European, or international road, a national, European, or international motorway, a forest path, a route for autonomous storage devices in a storage warehouse, etc.
[0006] A road consists of at least one carriageway. A carriageway is any physical means capable of supporting the movement of a vehicle. A motorway typically consists of two carriageways separated by a central reservation.
[0007] A roadway consists of at least one lane. A "lane" is any portion of a roadway designated for a single lane of vehicles. A motorway carriageway typically consists of at least two traffic lanes. Examples of lanes include a merging lane on a motorway, a single lane in a tunnel, a one-way traffic lane in a city, etc.
[0008] A lane may be delimited by road markings, but it may also correspond to a path on the roadway taken by vehicles traveling on the roadway. Such a lane may be called a "virtual lane" because this lane is not delimited by physical markings but is generated from past paths taken by vehicles traveling on the roadway.
[0009] Vehicles are often equipped with adaptive cruise control. Adaptive cruise control works by regulating the speed of the ego-vehicle based on, among other things, the speed of a target object (mostly a target vehicle) preceding the ego-vehicle in its lane.
[0010] Adaptive cruise control is also known as ACC, for Adaptive Cruise Control in English, the French translation being adaptive cruise control.
[0011] Selecting a target object for autonomous driving, such as adaptive cruise control, is crucial. In practice, situations where more than one target object is detected by the ego-vehicle are common.
[0012] For example, when the ego-vehicle is moving in the right-hand lane on a motorway in a country where driving is on the right, with two carriageways separated by a central reservation and each having two lanes (the right-hand lane and the left-hand lane), it is possible that two target objects, for example two other vehicles, are moving in front of the ego-vehicle.
[0013] Document US2014 / 330479 discloses predictive reasoning for regulating the speed of a vehicle. Document US2014 / 257665 discloses a method for regulating the speed and / or autonomy of motor vehicles. Document DE102007057722 discloses an adaptive cruise control system for a vehicle and a method for implementing the system. Document EP2562060 discloses a vehicle and a method for controlling the distance between moving vehicles.
[0014] Current autonomous driving methods usually handle this situation by using the target object closest to the ego vehicle as a reference for adaptive cruise control. With this method, the target object is correctly detected by the ego vehicle's sensors. However, this method presents significant safety deficiencies.
[0015] Indeed, it is possible that the object closest to the ego-vehicle exhibits dangerous behavior, for example by approaching very quickly a second target object with a lower speed. This situation is illustrated below with reference to the figure 1 .
[0016] In this situation, taking the dangerously driving object, here called the first target object, as a reference for adaptive cruise control is very risky. Indeed, the first target object can pull out very quickly to overtake the second target object in a brutal maneuver. This is not acceptable for the ego-vehicle, which may not have the same road-holding abilities, or simply because the passengers do not want to undergo such a maneuver.
[0017] The present invention improves the situation.
[0018] To this end, a first aspect of the invention relates to a method for autonomous driving of a vehicle, called ego-vehicle, in the presence of at least two target objects, a first target object and a second target object, according to claim 1.
[0019] Generating a driving instruction in which the speed of the ego-vehicle is regulated to the speed of the second target object while maintaining a target distance between the ego-vehicle and the first target object solves in a simple and comfortable manner for the occupants of the ego-vehicle the problems related to a plurality of vehicles traveling at different speeds in front of the ego-vehicle.
[0020] In fact, by regulating the speed on the second target object, if for example it has a speed much lower than that of the first target object, the risks linked to a sudden avoidance maneuver of this second target object are very reduced.
[0021] In addition, the distance between the ego-vehicle and the first target object is taken into account, so it remains possible to avoid any modification of the behavior of the first target object and the occupants of the ego-vehicle are not frightened by too close a proximity to this first target object.
[0022] A "target object" is any object present in the environment of the ego-vehicle. For example, any land vehicle located in front of (upstream of the trajectory of) the ego-vehicle, any type of obstacle, any disturbance on the roadway, etc. are target objects.
[0023] Furthermore, at least two target objects are mentioned here. The invention is of course not limited to two target objects but can be applied to an infinite number of target objects. Thus, for example, for three target objects, and as will be detailed below in the description, the speed of the first target object is used and weighted speeds are used for the other two target objects.
[0024] "Data" means any type of information, for example, available in digital form. The use of the plural form for "data" does not necessarily mean that more than one piece of data is present; for example, a bit can be understood as "data."
[0025] In one embodiment, the target speed of the ego-vehicle is longitudinal and corresponds to a speed of the second target object.
[0026] In one embodiment, the method further comprises a step of determining a type of the first target object and the target distance is further determined from said type of the first target object. By weighting the target distance by the type of the first target object, the dynamics, in particular, of the changes in the first target object are advantageously taken into account. Typically, a large object (truck for example) will have a greater movement inertia than a small object (small private vehicle for example).
[0027] In one embodiment, the type of the first target object is determined from said motion data of the first target object.
[0028] According to the invention, the method further comprises a consolidation of a presence indicator of the first target object and of the second target object and comprises the steps of: obtaining data from the sensor of the ego-vehicle, said data comprising information on the presence of the first target object and information on the presence of the second target object; estimating a trajectory of the ego-vehicle; calculating an instantaneous indicator of the presence of the first target object and an instantaneous indicator of the second target object from said sensor data and said estimated trajectory of the ego-vehicle; determining a consolidated indicator of the presence of the first target object and a consolidated indicator of the presence of the second target object from the instantaneous indicators of the presence of the first target object and the second target object and a history of consolidated indicators of the presence of the first target object and the second target object; determining a minimum between the consolidated indicator of the presence of the first target object and the consolidated indicator of the presence of the second target object; and the autonomous driving instruction is further generated from said minimum.
[0029] The processing of sensor data is therefore advantageously adapted to the specific functional requirements of driving an autonomous vehicle.
[0030] On the one hand, the consolidated presence indicator depends on the trajectory of the ego-vehicle. False off-trajectory detections are therefore filtered. In one embodiment, such filtering is continuous (progressive consideration depending on the proximity of the trajectory) or, in another embodiment, discrete (binary; consideration or not, or by discrete levels; given number of levels of consideration).
[0031] On the other hand, the presence indicator is consolidated by taking into account the history of the previous consolidated indicators. Therefore, all false detections (false positives) are smoothed by the history. For example, in one embodiment, the sudden appearance of an object at a given time without the object having ever been detected previously is discarded for the generation of the autonomous driving instruction.
[0032] The technical effects of filtering off-path detections and filtering previously undetected detections combine to have, in particular, a synergistic effect of introducing a consistency check. Indeed, only stable detection on the trajectory of the ego-vehicle guarantees that a relevant target is present. Thus, for example, all situations where a vehicle swerves slightly and periodically (typically to allow two wheels traveling between the lanes) into the lane of the ego-vehicle are filtered.
[0033] A second synergistic effect is linked to the appearance of stable off-trajectory detection errors. Indeed, many sensors in charge of autonomous driving are located facing the road, towards the front. Thus, during situations of projections on these sensors (rain, snow, mud, humidity, etc.), the drops will statistically concentrate outside the field of sensors corresponding to the trajectory. For example, the drops concentrate on the sides of the lens of a camera, the center (possibly off-center for significant lateral accelerations) being relatively protected from the drops by the apparent wind.
[0034] In one embodiment, the distance between the first target object and the ego-vehicle is less than the distance between the second target object and the ego-vehicle. In another embodiment, the first target object has a pre-collision time with the ego-vehicle that is less than a pre-collision time between the second target object and the ego-vehicle. In one embodiment, the longitudinal velocity of the second target object is less than or equal to the longitudinal velocity of the first target object.
[0035] In one embodiment, the autonomous driving instruction comprises an instruction for adaptive regulation of the speed of the ego-vehicle, the adaptive regulation being configured to take into account a virtual object moving at a speed linked to the target speed and / or at a distance from the ego-vehicle linked to the target distance. In one embodiment, linked means equal to. In another embodiment, linked means proportionally linked to.
[0036] A second aspect of the invention relates to a computer program comprising instructions for implementing the method according to the first aspect of the invention, when these instructions are executed by a processor.
[0037] A third aspect of the invention relates to a device for autonomous driving of a vehicle, called an ego-vehicle, according to claim 8.
[0038] A fourth aspect of the invention relates to a vehicle comprising the device according to the third aspect of the invention.
[0039] Other characteristics and advantages of the invention will appear on examining the detailed description below, and the appended drawings in which: [ Fig.1 ] illustrates a context of application of the invention; [ Fig.2 ] illustrates a method according to the invention; [ Fig.3 ] illustrates a device, according to one embodiment of the invention.
[0040] The invention is described below in its non-limiting application to the case of an autonomous motor vehicle traveling on a highway comprising two carriageways separated by a central reservation and each comprising two lanes. Other applications such as a bus on a dedicated lane or a motorcycle on a country road are also conceivable.
[0041] There figure 1 illustrates a motorway carriageway with two lanes, a left lane L and a right lane R with the same direction of travel (from the bottom to the top of the figure). The left lane L has an ego-vehicle EV, a second target object VCI and a first target object VC. Here, the distance between the first target object and the ego-vehicle is less than the distance between the second target object and the ego-vehicle.
[0042] However, in another embodiment, to determine the first target object and the second target object, the following rule is used: the first target object has a time-to-collision with the ego-vehicle that is less than a time-to-collision between the second target object and the ego-vehicle. A time-to-collision (TTC) is any type of duration corresponding to the time, at a time t, that the ego-vehicle would take to hit a target object if the motion data of the ego-vehicle and the target object remain those present at time t.
[0043] In another embodiment, the longitudinal velocity of the second target object is less than or equal to the longitudinal velocity of the first target object.
[0044] The first target object VC changes lanes according to an INS maneuver. The new VC lane here is a motorway exit VS. Of course, other final lanes for VC can be chosen, the rightmost lane on the motorway, an emergency braking lane, etc.
[0045] To illustrate the relevance of the method according to the invention, described below in detail with reference to the figure 2 , it can be considered that the first target object VC has a speed much higher than that of the second target object VCI and does not brake in view of its exit from the motorway via the VS exit.
[0046] This situation is relatively common in urban environments, typically on the ring roads of large cities, where many drivers drive aggressively. In such a situation, the adaptive cruise control of the EV must anticipate the second VCI target object as best as possible.
[0047] There figure 2 illustrates the method, according to one embodiment of the invention.
[0048] In a step 10, a sensor C of the ego-vehicle acquires motion data of the first target object VC and motion data of the second target object VCI.
[0049] The ego-vehicle's sensor C is at least one of the following: a radar; a lidar; a camera, for example a multifunction video camera, CVM; a communication system (for example Car2X or 5G) configured to receive information from at least one other vehicle, infrastructure, user terminal, etc.; a laser; etc.
[0050] Raw data acquired by sensor C is processed and merged when multiple sensors are used.
[0051] It is here confirmed that the sensors of the ego-vehicle EV can indeed, in certain situations, acquire data on the second target object VCI, even if it is partially masked by the first target object VC. For example, the waves emitted by a radar pass under the first target object VC to locate the second target object. In another example, the movement data of VCI is directly transmitted by 5G communication to the ego-vehicle EV.
[0052] In an embodiment comprising a type, a type of the first target object VC is also acquired. A target object type corresponds to any type of target object category. For example, a target object type corresponds to an object category (private vehicle, truck, static obstacle, animal, motorcycle, pedestrian, etc.), an importance classification (security vehicle, personality vehicle, etc.), etc.
[0053] In a step 12, a target speed V_VCI of the ego-vehicle from the motion data of the second target object. In one embodiment, the target speed of the ego-vehicle is longitudinal and corresponds to a speed of the second target object.
[0054] In a step 14, a target distance ΔD_VC_EV between the ego-vehicle and the first target object from the motion data of the first target object. In the embodiment comprising the type, the target distance is further determined from the type of the first target object.
[0055] In a step 16, an autonomous driving instruction of the ego-vehicle is generated from the target speed V_VCI and the target distance ΔD_VC_EV. In particular, in one embodiment, the autonomous driving instruction comprises an instruction for adaptive regulation of the speed of the ego-vehicle, the adaptive regulation being configured to take into account a virtual object moving at a speed V_EV linked to the target speed and / or at a distance ΔD from the ego-vehicle linked to the target distance ΔD_VC_EV.
[0056] According to the invention, the method further comprises a consolidation of a presence indicator of the first target object and of the second target object and comprises the steps of: obtaining data from the sensor of the ego-vehicle, said data comprising information on the presence of the first target object and information on the presence of the second target object; estimating a trajectory of the ego-vehicle.In particular, the trajectory of the ego-vehicle can be estimated from movement data of the ego-vehicle, position data of a road on which said ego-vehicle is moving and position data of a lane included on said road; calculation of an instantaneous indicator of presence of the first target object and an instantaneous indicator of the second target object from said sensor data and said estimated trajectory of the ego-vehicle; determination of a consolidated presence indicator of the first target object and a consolidated presence indicator of the second target object from the instantaneous indicators of presence of the first target object and the second target object and a history of consolidated presence indicators of the first target object and the second target object.In particular, the instantaneous presence indicator is calculated at a given time (t), and in which the consolidated presence indicator is determined from the instantaneous presence indicator at the given time and the history of consolidated presence indicators comprising at least one consolidated presence indicator at a time preceding (t-1) the given time; determination of a minimum between the consolidated presence indicator of the first target object and the consolidated presence indicator of the second target object; . and the autonomous driving instruction is further generated from said minimum.
[0057] A "presence indicator" means any data characteristic of the presence of an object. In particular, at least one numerical value for which a large value indicates a high probability of the object's presence and a small value indicates a low probability of the object's presence is an example of a presence indicator.
[0058] There figure 3 represents an example of a device D included in the ego-vehicle EV. This device D can be used as a centralized device in charge of at least certain steps of the method described above with reference to the figure 2 .
[0059] This device D can take the form of a box containing printed circuits, any type of computer or even a smartphone.
[0060] The device D comprises a RAM 1 for storing instructions for the implementation by a processor 2 of at least one step of the methods as described above. The device also comprises a mass memory 3 for the storage of data intended to be retained after the implementation of the method.
[0061] The device D may further comprise a digital signal processor (DSP) 4. This DSP 4 receives data to format, demodulate and amplify this data in a manner known per se.
[0062] The device also comprises an input interface 5 for receiving data implemented by methods according to the invention and an output interface 6 for transmitting data implemented by the methods.
[0063] The present invention is not limited to the embodiments described above as examples; it extends to other variants according to the appended claims.
[0064] Thus, an exemplary embodiment has been described above in which the ego-vehicle was moving on a motorway comprising two carriageways of two lanes, the carriageways being separated by a central reservation. The invention can also be implemented for other types of roads, such as national roads with a single carriageway comprising two lanes, motorways with separate carriageways, each carriageway comprising 6 lanes, etc.
[0065] In addition, an embodiment has been described in which the invention was described in the context of the application to adaptive speed regulation.
[0066] In addition, the steps described with reference to the figure 2 were carried out in a specific order. A different order is also possible. For example, steps 12 and 14 can be reversed or carried out simultaneously.
Claims
1. Method for autonomous driving of a vehicle, called an ego-vehicle (EV), in the presence of at least two target objects, a first target object and a second target object, the method comprising the steps of: - acquisition (10) by at least one sensor of the ego-vehicle of movement data of the first target object; - acquisition (10) by at least one sensor of the ego-vehicle of movement data of the second target object; - determination of a target speed of the ego-vehicle from the movement data of the second target object; - determination of a target distance between the ego-vehicle and the first target object from the movement data of the first target object; - consolidation of an indicator of presence of the first target object and of the second target object, characterized in that the method comprises the steps of: • obtaining data from the sensor of the ego-vehicle, said data comprising information on the presence of the first target object and information on the presence of the second target object; • estimation of a trajectory of the ego-vehicle; • calculation of an instantaneous indicator of presence of the first target object and an instantaneous indicator of the second target object from said sensor data and said estimated trajectory of the ego-vehicle; • determination of a consolidated indicator of presence of the first target object and a consolidated indicator of presence of the second target object from the instantaneous indicators of presence of the first target object and the second target object and a history of consolidated indicators of presence of the first target object and the second target object; • determination of a minimum between the consolidated indicator of presence of the first target object and the consolidated indicator of presence of the second target object; - generation of an autonomous driving instruction of the ego-vehicle from the target speed, the target distance, and said minimum.
2. The method according to claim 1, wherein the target speed of the ego-vehicle is longitudinal and corresponds to a speed of the second target object.
3. Method according to one of the preceding claims, further comprising a step of: - determining a type of the first target object; and wherein the target distance is further determined from said type of the first target object.
4. Method according to claim 3, wherein the type of the first target object is determined from said motion data of the first target object.
5. Method according to claim 1, wherein the distance between the first target object and the ego-vehicle is less than the distance between the second target object and the ego-vehicle.
6. Method according to claim 1, wherein the first target object has a time before collision with the ego-vehicle less than a time before collision between the second target object and the ego-vehicle.
7. Computer program comprising instructions for implementing the method according to any one of the preceding claims, when these instructions are executed by a processor (2).
8. Device (D) for autonomous driving of a vehicle, called an ego-vehicle (EV), in the presence of at least two target objects, a first target object and a second target object, the device comprising: - at least one sensor (C) configured for acquiring movement data of the first target object and movement data of the second target object; - at least one processor (2) and at least one memory (1, 3) arranged to perform the operations of: ∘ determining a target speed of the ego-vehicle from the movement data of the second target object; ∘ determining a target distance between the ego-vehicle and the first target object from the movement data of the first target object; ∘ consolidation of a presence indicator of the first target object and of the second target object, comprising the steps of: • obtaining data from the sensor of the ego-vehicle, said data comprising information on the presence of the first target object and information on the presence of the second target object; • estimating a trajectory of the ego-vehicle; • calculating an instantaneous indicator of the presence of the first target object and an instantaneous indicator of the second target object from said sensor data and said estimated trajectory of the ego-vehicle; • determining a consolidated presence indicator of the first target object and a consolidated presence indicator of the second target object from the instantaneous indicators of the presence of the first target object and of the second target object and from a history of consolidated presence indicators of the first target object and of the second target object; • determining a minimum between the consolidated presence indicator of the first target object and the consolidated presence indicator of the second target object; ∘ generation of an autonomous driving instruction of the ego-vehicle from the target speed, the target distance, and said minimum.
9. Vehicle (EV) comprising the device according to claim 8.
