JP6213277B2 - Vehicle control apparatus and program - Google Patents

Description

  The present invention relates to a vehicle control device and a program, and more particularly to a vehicle control device for realizing optimal vehicle control.

  Conventionally, as a vehicle control method on a road, a method is known in which road information is acquired in advance and a vehicle is controlled by selecting which road to travel (Non-Patent Document 1). For example, a selection is made as to whether the vehicle is running on the road that is currently running or whether the lane is changed to the adjacent lane.

  In addition, unmanned aircraft in team formation can be used as a method for efficiently searching for designated areas in order to search for targets to be bombed. There is known a method of expressing the behavior that can be taken (Non-Patent Document 2). The evaluation function used in this method is how much the specified area can be searched, and the evaluation becomes higher if the search is exhaustively performed. For this reason, the problem of finding an optimal solution from a combination of discretized actions is a problem that increases exponentially and cannot be calculated, so the problem can be solved by applying the DP algorithm. doing.

Michael Montemerlo, Jan Becker, Suhrid Bhat, Hendrik Dahlkamp and Dmitri Dolgov, Scott Ettinger, Dirk Haehnel, `` Junior: The Stanford Entry in the Urban Challenge '', Jornal of Field Robotics, 2008, 25 (9), 569-597 Matthew Flint, Marios Polycarpou, Emmanuel Fernandez-Gaucherand, "COOPERATIVE PATH-PLANNING FOR AUTONOMOUS VEHICLES USING DYNAMIC PROGRAMMING", 15th Triennial World Congress, Barcelona, Spain, 2002

  However, in the method of Non-Patent Document 1, there are only two selection candidates that can be broadly divided into whether the road is currently running or whether the lane is changed to the adjacent lane. Since the behavior of changing the lane after driving a little straight is not considered, the course is determined based only on the current state information.

  Considering the case where another vehicle crosses the road as shown in FIG. 7, when the course is determined only in the current state, the other lane exists in the current state in the adjacent lane, so the lane is changed to the adjacent lane. The course to be performed is determined to be impossible. For this reason, the host vehicle tries to move forward as it is, but at the next time, the other vehicle moves forward and enters the traveling lane of the host vehicle. At that time, if there is no room for the vehicle to change lanes, there is a problem that it is necessary to decelerate rapidly. Sudden deceleration is a cause that adversely affects everything in terms of fuel consumption, arrival time, and riding comfort, and should be reduced as much as possible.

  Further, in the method of Non-Patent Document 2, since the controlled object moves at a constant speed, it becomes possible to divide the combination search problem into partial problems, and as a result, the DP algorithm is applied. However, in the method of Non-Patent Document 2, when the speed of the control target is variable, the combination search problem cannot be divided into partial problems, and the DP algorithm cannot be applied by this method. There is.

  An object of the present invention is to provide a vehicle control device that can search for a time-series pattern of an optimal vehicle state of a host vehicle until a predetermined time after allowing a change in speed.

  In order to achieve the above object, a vehicle control device according to a first aspect of the present invention includes vehicle information acquisition means for acquiring vehicle information including position information for each of the host vehicle and other vehicles, an acceleration state, a deceleration state, and a constant speed. A node representing a plurality of types of vehicle states including a state and a lane change state for each time up to a predetermined time, and connecting the nodes in time series, the time series of the vehicle state of the host vehicle In the vehicle state network representing each of the patterns, for each node, the node obtained based on the vehicle state represented by the node and the vehicle information of each of the own vehicle and other vehicles acquired by the vehicle information acquisition means The time series pattern of the optimal vehicle state of the host vehicle is calculated according to the cost calculated based on the speed and acceleration of the host vehicle at the time corresponding to Includes a optimal solution search means search, the have been configured.

  According to the first invention, the vehicle information acquisition unit acquires vehicle information including position information for each of the host vehicle and the other vehicle, and the optimum solution search unit acquires the acceleration state, the deceleration state, the constant speed state, and Each of the nodes representing a plurality of types of vehicle states including a lane change state is provided for each time up to a predetermined time, and each node is connected in time series to represent each time series pattern of the vehicle state of the host vehicle. In the vehicle state network, the time series pattern of the optimal vehicle state of the host vehicle is searched for each node according to the calculated cost.

  As described above, each node has a plurality of types of vehicle states including an acceleration state, a deceleration state, a constant speed state, and a lane change state for each time until a predetermined time, and the nodes are connected in time series. In the vehicle state network representing each of the time series patterns of the vehicle state of the host vehicle, the speed of the vehicle is determined by searching the time series pattern of the optimum vehicle state of the host vehicle according to the calculated cost for each node. The time series pattern of the optimal vehicle state of the host vehicle until a predetermined time can be searched while allowing the change.

  In the first invention, the optimum solution search means is a constraint condition that includes the lane change state only once in the time series pattern of the optimum vehicle state of the host vehicle, and the optimal vehicle state of the host vehicle. Constraints that do not transition to the deceleration state immediately after the acceleration state in the series pattern, constraints that do not transition to the acceleration state immediately after the deceleration state in the time series pattern of the optimal vehicle state of the host vehicle, and At least one of the constraints that the time series pattern of the optimal vehicle state of the host vehicle is searched according to the result of evaluating the cost by an A * algorithm using a heuristic function formula determined according to the formula for calculating the cost You may make it search the time series pattern of the optimal vehicle state of the own vehicle so that constraint conditions may be satisfy | filled.

  In the first invention, the optimum solution search means uses the cost of the node calculated according to a predetermined formula including a fuel consumption term, a speed limit term, and a safe inter-vehicle term, to determine the optimum vehicle's optimum. You may make it search the time series pattern of a vehicle state.

  In the first invention, the control means for controlling the driving of the host vehicle based on the vehicle state one hour later in the time series pattern of the optimum vehicle state of the host vehicle searched by the optimum solution searching unit. May further be included.

  Moreover, the program of this invention is a program for functioning a computer as each means which comprises said vehicle control apparatus.

  As described above, according to the vehicle control device and the program of the present invention, each of the nodes representing a plurality of types of vehicle states including the acceleration state, the deceleration state, the constant speed state, and the lane change state is set until a predetermined time. In the vehicle state network that represents each of the time-series patterns of the vehicle state of the host vehicle by connecting the nodes in time series, the optimal state of the host vehicle is determined according to the calculated cost for each node. By searching for the time series pattern of the vehicle state, it is possible to allow a change in speed and to search for the time series pattern of the optimal vehicle state of the host vehicle until a predetermined time later.

It is a figure which shows the example of a discrete state. It is a figure which shows the example of a vehicle state network. It is a figure which shows the example which calculated the cost of the vehicle state node. It is a block diagram which shows the functional structure of the vehicle control apparatus which concerns on embodiment of this invention. It is a flowchart figure which shows the vehicle control processing routine in the vehicle control apparatus which concerns on embodiment of this invention. It is a figure which shows the experiment example using the vehicle control apparatus which concerns on embodiment of this invention. It is a figure which shows the example which another vehicle crosses a road.

  Embodiments of the present invention will be described below with reference to the drawings. In the drawings, substantially the same or equivalent components or parts are denoted by the same reference numerals.

<Principle of the invention>
In the embodiment of the present invention, first, the state of the host vehicle is discretized. For example, as shown in FIG. 1, (A) acceleration state, (D) deceleration state, (C) constant speed state, and (L) lane change state are divided into four states, and each of the divided states is a vehicle state. Let it be a node. In the acceleration state and the deceleration state, a predetermined constant acceleration or deceleration is performed. The vehicle state node is an example of a node. Four types are examples of a plurality of types.

Next, the four divided vehicle state nodes are provided for each time up to an arbitrary time, and are connected in time series. In the present embodiment, the predicted time interval is set to 1 second. In this case, when considering the time series pattern of the vehicle state up to one second later, there are four time series patterns of the vehicle state in all of (A), (D), (C), and (L). It will be. When considering the time-series pattern of the vehicle state up to 2 seconds later, (A)-(A), (A)-(D), (A)-(C), (A)-(L), ( D)-(A),..., (L)-(C), (L)-(L), there are 16 time series patterns of vehicle states. Therefore, in consideration of the time series pattern of the vehicle state up to an arbitrary time n seconds later, a vehicle state network represented by a time series pattern of 4 ⁿ vehicle states is configured. FIG. 2 shows an example of a vehicle state network.

  And in each vehicle state node in the created vehicle state network, the cost in the vehicle state node is calculated. As a cost calculation method, for example, when performing vehicle control with optimized fuel efficiency, the following formula (1) is used. The sum of the costs in the final state is obtained by adding the value of the following equation (1) and the sum of the costs up to the previous state.

Where v is the speed of the host vehicle at the time corresponding to the vehicle state node, a is the acceleration of the host vehicle at the time corresponding to the vehicle state node, and x _near corresponds to the vehicle state node. Is the distance to the other vehicle closest to the own vehicle in the traveling direction at the time to be performed, and y _near is the distance to the other vehicle closest to the own vehicle in the direction perpendicular to the traveling direction at the time corresponding to the vehicle state node. Is the speed limit of the road on which the host vehicle is located, t is the predicted time interval (seconds), W is a function related to fuel consumption, and λ (λ _a is an adjustment parameter of the fuel consumption term, λ _{b is} an adjustment parameter of a speed limit term, λ _{c is} an adjustment parameter of a safe inter-vehicle term.), α, and β are predetermined adjustment parameters. In the present embodiment, t is 1. Further, the first term of the above equation (1) is the fuel consumption term, the second term is the speed limit term, and the third term is the safe inter-vehicle term. FIG. 3 shows an example of cost calculation performed on the vehicle state network of FIG. Note that x _near and y _near are examples of relative positions with respect to other vehicles. The predicted time interval is an example of a time interval corresponding to a node.

  Then, a time series pattern of an optimal vehicle state is searched for in a cost-added network as shown in FIG. The time series pattern of the optimal vehicle state in the predictive vehicle control is a time series pattern in which the sum of costs at a certain arbitrary time is minimized. As a network search method, for example, a conventional technique such as the Dijkstra method may be used.

  In the network shown in FIG. 3, since the sum of costs up to the vehicle state node (L) is minimum at the time after 4 seconds, the time series of (C)-(A)-(A)-(L) The pattern is determined to be optimal. This time-series pattern is a time-series pattern in which acceleration control is performed for 2 seconds after the vehicle is controlled at a constant speed for 1 second, and then lane change control is performed. In actual vehicle control, only the control method corresponding to the vehicle state node after one hour is used. In the specific example, the constant speed control in (C) is commanded to the host vehicle. Note that 4 seconds later is an example after a predetermined time.

  Although it is possible to realize predictive vehicle control by using the above method, there is a problem of calculation amount. Since the vehicle state network is expressed by a combination of discrete states, the longer the time to be predicted, the more the combination explosion occurs and the processing cannot be calculated in real time. When the number of discrete states of the vehicle is S and the length of time to be predicted is t, the calculation amount order is O (S ′). Therefore, in the present embodiment, the number of searches is reduced using the following four constraints. It should be noted that the constraint condition (2) below does not transition to the deceleration state immediately after the acceleration state, and the constraint condition (3) below is immediately after the deceleration state in the time series pattern of the optimal vehicle state of the host vehicle. This is an example of a constraint condition that the transition to the acceleration state is not performed.

(1) Restriction condition that the lane change state is included only once in the state series as the optimal solution (2) Restriction condition that the state series as the optimal solution does not decelerate immediately after acceleration (3) Optimal solution The state series to be expressed as follows: Constraint condition that acceleration is not performed immediately after deceleration (4) A value calculated by a heuristic function equation (equation (2) below) determined according to equation (1) above, * Restriction condition to search the time series pattern of the optimal vehicle state of the vehicle according to the result of evaluating the cost of the vehicle state node by the algorithm

ただし、ｖは、当該車両状態ノードが表わす車両状態に基づいて求められる、当該車両状態ノードに対応する時刻における自車両の速度であり、ｌは、自車両が位置する道路の制限速度である。

Where v is the speed of the host vehicle at the time corresponding to the vehicle state node, which is obtained based on the vehicle state represented by the vehicle state node, and l is the speed limit of the road on which the host vehicle is located.

<Configuration of Vehicle Control Device according to Embodiment of the Present Invention>
Next, the vehicle control apparatus according to the embodiment of the present invention will be described. As shown in FIG. 4, the vehicle control device 100 according to the embodiment of the present invention includes a communication unit 10, a GPS 12, a speed sensor 14, an acceleration sensor 16, a calculation unit 20, and a driving support unit 50. I have.

  The communication unit 10 receives time-series data of position information of the other vehicle from each of the other vehicles located within a certain range from the host vehicle, and outputs the time series data to the calculation unit 20.

  The GPS 12 acquires the GPS information of the host vehicle and outputs it to the calculation unit 20.

  The speed sensor 14 measures the speed of the host vehicle and outputs it to the calculation unit 20.

  The acceleration sensor 16 measures the acceleration of the host vehicle and outputs it to the calculation unit 20.

  The calculation unit 20 includes a vehicle information acquisition unit 22, an optimum solution search unit 24, a control unit 26, and a vehicle state network storage unit 28.

Vehicle state network storage unit 28 has been pre-built stores vehicle state network indicating a vehicle state of the vehicle of 4 ⁿ street prediction time interval until the n seconds after the t seconds. The vehicle state network storage unit 28 stores adjustment parameters λ, α, and β corresponding to the above equation (1). In the present embodiment, the vehicle state network, a vehicle state network indicating a vehicle state of the vehicle 4 are ^four up to 4 seconds after a 1 second prediction time interval.

  The vehicle information acquisition unit 22 acquires position information about each of the other vehicles located within a certain range from the host vehicle input from the communication unit 10 in time series. Moreover, the vehicle information acquisition part 22 acquires the position of the own vehicle based on GPS information input from GPS12, and acquires the road information containing the speed limit information of the road where the own vehicle is located from road map data. . Further, the vehicle information acquisition unit 22 acquires the speed of the host vehicle input from the speed sensor 14 and the acceleration of the host vehicle input from the acceleration sensor 16. The position information acquired in time series for each of the other vehicles is an example of the vehicle information of the other vehicle, and the position of the host vehicle, road information, the speed of the host vehicle, and the acceleration of the host vehicle are It is an example of the vehicle information of the own vehicle.

  The optimal solution search unit 24 acquires the vehicle state network stored in the vehicle state network storage unit 28, and performs Dijkstra method for the vehicle state network according to the cost calculated for each vehicle state node and the four constraints. The time series pattern of the optimal vehicle state that satisfies the four constraint conditions and minimizes the total cost is searched for and output to the control unit 26.

  Specifically, for the vehicle state network, the cost of the vehicle state node is calculated while searching for each vehicle state node according to the Dijkstra method and the four constraints (1) to (4) above, and the sum of the costs is calculated. The time series pattern of the vehicle state that minimizes is searched.

  Here, each information of the above equation (1) used when calculating the cost of the vehicle state node includes the adjustment parameters (λ, α, and β) stored in the vehicle state network storage unit 28, the vehicle The speed, acceleration, position, road information of the host vehicle acquired in the information acquisition unit 22, or the position, speed, and acceleration of the host vehicle in the vehicle state node immediately before the vehicle state node, and the vehicle information acquisition unit 22 Is obtained from the position information of each of the other vehicles acquired in step 1 and the vehicle state of the vehicle state node. For each of the other vehicles, the current speed and the moving direction of each other vehicle are calculated from the position information acquired in time series, and the calculated moving direction is calculated at a constant speed. Is assumed to move every second.

  Specifically, the speed v of the host vehicle is obtained based on the speed of the host vehicle at the vehicle state node one hour before, and when the vehicle state node one hour before is in the acceleration state, only the predetermined speed is obtained. When the vehicle state node at one time ago is in a decelerating state, the speed is obtained by subtracting a predetermined speed. The acceleration a of the host vehicle is a positive predetermined acceleration when the vehicle state node is in an acceleration state, and a negative predetermined acceleration when the vehicle state node is in a deceleration state. Further, each distance to the other vehicle closest to the host vehicle in the traveling direction and in the direction perpendicular to the traveling direction is obtained as follows. First, based on the position information, the speed, and the acceleration of the own vehicle acquired by the vehicle information acquisition unit 22, or the position information of the own vehicle obtained for the vehicle state node one hour ago, and one hour before Based on the speed of the host vehicle calculated with respect to the vehicle state node and the acceleration of the host vehicle calculated with respect to the vehicle state node one hour before, the position of the host vehicle at the time of the vehicle state node is calculated. Is done. If the vehicle state node at one hour before is a lane change, the position of the host vehicle after the lane change is calculated. Further, the position information of the other vehicle acquired by the vehicle information acquisition unit 22, or the position information of the other vehicle determined with respect to the vehicle state node one time ago and the time series of the position information of the other vehicle. Based on the speed and the moving direction of the vehicle, the position of the other vehicle at the time of the vehicle state node is calculated. As described above, the position of the own vehicle and the position of the other vehicle at the time of the vehicle state node are calculated, and based on the calculated result, the position of the own vehicle in the traveling direction and in the direction perpendicular to the traveling direction is the most. Each distance to a nearby other vehicle is determined. Further, the road speed limit is obtained from the road information acquired by the vehicle information acquisition unit 22.

  The control unit 26 outputs a vehicle control command to the driving support unit 50 based on the time-series pattern of the vehicle state that satisfies the constraint condition and that minimizes the total cost, searched for by the optimal solution search unit 24. To do. Specifically, when the optimal solution search unit 24 searches for a time series pattern ((C)-(A)-(A)-(L)) of the vehicle state as shown in FIG. A command for constant speed control corresponding to the vehicle state is output to the driving support unit 50. The time-series pattern of the vehicle state is one in which the constant speed control is performed for 1 second, the acceleration control is performed for 2 seconds, and then the lane change control is performed.

  The driving support unit 50 performs automatic driving control on the host vehicle in accordance with a command input from the control unit 26.

<Operation of the vehicle control device according to the present embodiment>
Next, the operation of the vehicle control device 100 according to the embodiment of the present invention will be described. First, the communication unit 10 sequentially receives position information of each of the other vehicles located within a certain range from the own vehicle, the GPS 12 sequentially detects the GPS information of the own vehicle, and the speed sensor 14 determines the speed of the own vehicle. The vehicle control processing routine shown in FIG. 5 is executed by the CPU executing the program stored in the ROM of the vehicle control device 100 when the acceleration sensor 16 detects the acceleration of the host vehicle sequentially. Is executed.

  First, in step S100, the vehicle information acquisition unit 22 detects each position information of each other vehicle received in time series by the communication unit 10, the GPS information of the own vehicle detected by the GPS 12, and the speed sensor 14. The speed of the host vehicle and the acceleration of the host vehicle detected by the acceleration sensor 16 are received.

  In step S101, the position of the host vehicle is calculated from the GPS information of the host vehicle acquired in step S100, and road information including speed limit information of the road on which the host vehicle is located is acquired from the road map data.

  Next, in step S102, the vehicle state network stored in the vehicle state network storage unit 28 and the adjustment parameters λ, α, and β are read.

  Next, in step S104, each position information of each other vehicle acquired in step S100, the speed of the host vehicle, the acceleration of the host vehicle, the position and road information of the host vehicle acquired in step S101, Based on the adjustment parameters λ, α, and β acquired in step S102, for the vehicle state network acquired in step S102, for the vehicle state network, the cost calculated according to the above equation (1) for each vehicle state node, In accordance with the four constraints (1) to (4) above, the Dijkstra method is used to search for a time series pattern of the optimal vehicle state that satisfies the four constraints and minimizes the total cost. .

  Next, in step S106, a command for controlling the vehicle is output to the driving support unit 50 based on the vehicle state one hour later in the time series pattern of the optimal vehicle state acquired in step S104.

  Next, in step S108, the automatic driving of the host vehicle is controlled according to the control command acquired in step S106, the process proceeds to step S100, and the processes in steps S100 to S108 are repeated. Note that the iterative process is performed at each predicted time interval.

<Experimental example>
When the vehicle control device 100 according to the present embodiment and the vehicle control device that determines the course using only the current situation in the prior art are compared using a simulation, the fuel efficiency of the vehicle control device 100 according to the present embodiment is improved. I was able to confirm. Here, the condition of this experimental example is a highway including a lane reduction scene, and the fuel consumption of a passing vehicle is measured by monitoring the section for one hour, and the prediction is made for prediction up to 5 seconds ahead. .

  FIG. 6 shows the number of searches when the four constraint conditions (1) to (4) are introduced. As shown in FIG. 6, the amount of calculation can be reduced by introducing the four constraint conditions (1) to (4).

  As described above, according to the vehicle control device 100 according to the present embodiment, each of the vehicle state nodes representing the four types of vehicle states including the acceleration state, the deceleration state, the constant speed state, and the lane change state is displayed. , Each vehicle state node is calculated for each vehicle state node in a vehicle state network having each time until a predetermined time and connecting each vehicle state node in time series to represent each of the time series patterns of the vehicle state of the host vehicle. By searching for the time series pattern of the optimal vehicle state of the host vehicle according to the cost, it is possible to search for the time series pattern of the optimal vehicle state of the host vehicle up to a predetermined time after allowing a change in speed. it can.

  Further, by using the vehicle control device 100 according to the present embodiment, it is possible to realize optimal vehicle control until after a predetermined time. By realizing optimal vehicle control up to a predetermined time in this way, higher-performance automatic driving control can be realized. For example, automatic driving with higher fuel consumption, safer automatic driving, and automatic driving with less shaking can be realized. Furthermore, it is also possible to process such predictive vehicle control in a realistic calculation time.

  In addition, the vehicle behavior up to a predetermined time is inherently infinite, but by dividing the vehicle state into a plurality of discrete states, the possibility of the vehicle behavior up to the predetermined time can be reduced. Processing in realistic calculation time can be realized. Furthermore, the processing speed can be increased by reducing the number of searches using four constraint conditions using driving behavior knowledge.

  In addition, for combinatorial optimization problems that cannot be solved by the DP algorithm, it is possible to derive an optimal solution by using the setting of a state sequence that allows a change in speed and the reduction of the number of searches using knowledge. .

  Note that the present invention is not limited to the above-described embodiment, and various modifications and applications are possible without departing from the gist of the present invention.

  In the present embodiment, the case where the position information of each other vehicle is acquired in time series by the communication unit 10 has been described, but the present invention is not limited to this. For example, sensors (cameras, laser radars, millimeter wave radars, etc.) that can recognize the environment are mounted on the host vehicle, and information on other surrounding vehicles (position, speed, and acceleration) is acquired by using the sensor. May be. Moreover, you may acquire each GPS information of another vehicle by the communication part 10 in time series. In addition, a configuration capable of acquiring the position, speed, and acceleration of another vehicle may be used.

In the present embodiment, the case where the discrete state is set to four types of vehicle states, that is, an acceleration state, a deceleration state, a constant speed state, and a lane change state, is not limited to this. For example, four or more types of discrete states including the four types of vehicle states may be used. In addition, a plurality of acceleration states may be used as the acceleration state, and a plurality of deceleration states may be used as the deceleration state. For example, a first acceleration state that accelerates by 10 m / s ² and a second acceleration state that accelerates by 20 m / s ² may be used as the acceleration state.

  Moreover, in this Embodiment, although the value calculated according to said Formula (1) was defined as a cost of a vehicle state node, it is not limited to this. For example, the sum of the value calculated according to the above equation (1) for the vehicle state node and the cost of the vehicle state node immediately before the vehicle state node may be defined as the cost of the vehicle state node.

  Further, in the present embodiment, the case has been described where all of the four constraints (1) to (4) are used, but the present invention is not limited to this. For example, at least one constraint condition among the four constraint conditions (1) to (4) may be used.

  In the present embodiment, the case where the speed limit of the road is obtained from the road information acquired by the vehicle information acquisition unit 22 has been described, but the present invention is not limited to this. For example, the speed limit of the road may be obtained from the position of the host vehicle at the vehicle node and the road information acquired by the vehicle information acquisition unit 22.

  Further, in the present specification, the embodiment has been described in which the program is installed in advance. However, the program can be provided by being stored in a computer-readable recording medium or provided via a network. It is also possible to do.

10 Communication unit 12 GPS
14 Speed sensor 16 Acceleration sensor 20 Calculation unit 22 Vehicle information acquisition unit 24 Optimal solution search unit 26 Control unit 28 Vehicle state network storage unit 50 Driving support unit 100 Vehicle control device

Claims (7)

Vehicle information acquisition means for acquiring vehicle information including position information for each of the host vehicle and other vehicles;
Each of the nodes representing a plurality of types of vehicle states including an acceleration state, a deceleration state, a constant speed state, and a lane change state for each time until a predetermined time, and connecting each node in time series, In the vehicle state network representing each time-series pattern of the vehicle state of the host vehicle, for each node, the vehicle state represented by the node and the vehicle information of each of the host vehicle and other vehicles acquired by the vehicle information acquisition unit Optimum to search for a time series pattern of the optimal vehicle state of the host vehicle according to the cost calculated based on the speed, acceleration and relative position of the host vehicle at the time corresponding to the node obtained based on Solution search means;
A vehicle control apparatus.   The optimal solution searching means includes a constraint condition that includes the lane change state only once in the time series pattern of the optimum vehicle state of the host vehicle, and immediately after the acceleration state in the time series pattern of the optimum vehicle state of the host vehicle. According to the constraint condition that the vehicle does not transition to the deceleration state, the constraint condition that the vehicle does not transition to the acceleration state immediately after the deceleration state in the time series pattern of the optimal vehicle state of the host vehicle, and the calculation formula for the cost In order to satisfy at least one of the constraints of searching for a time series pattern of the optimal vehicle state of the host vehicle according to the result of evaluating the cost by an A * algorithm using a heuristic function formula The vehicle control device according to claim 1, wherein a time series pattern of an optimal vehicle state of the vehicle is searched.   The optimum solution searching means searches for a time series pattern of an optimal vehicle state of the host vehicle using the cost of the node calculated according to a predetermined formula including a fuel consumption term, a speed limit term, and a safe inter-vehicle term. The vehicle control device according to claim 1 or 2. 4. The vehicle control device according to claim 3, wherein the optimum solution search means searches for a time series pattern of an optimum vehicle state of the host vehicle using the cost of the node calculated according to the following equation (1). .

However, v is the speed of the own vehicle at the time corresponding to the node, which is obtained based on the vehicle state represented by the node one hour before, and a is obtained based on the vehicle state represented by the node. The acceleration of the host vehicle at the time corresponding to the node, and x _near is the vehicle information of the host vehicle and other vehicles acquired by the vehicle information acquisition unit and the vehicle state represented by the node one hour before. Calculated based on the distance to the other vehicle closest to the own vehicle in the traveling direction at the time corresponding to the node, y _near is each of the own vehicle and the other vehicle acquired by the vehicle information acquisition means The vehicle in the direction perpendicular to the traveling direction at the time corresponding to the node, which is obtained based on the vehicle information and the vehicle state represented by the node one hour before. Is a distance to a nearby vehicle, l is a road speed limit obtained based on the position information of the host vehicle acquired by the vehicle information acquisition means, and t is a time interval corresponding to the node. W is a function related to fuel consumption, and λ _a , λ _b , λ _c , α, and β are predetermined adjustment parameters. The optimum solution searching means uses a value calculated by the following equation (2), and according to the result of evaluating the cost of the node by the A * algorithm, the time series pattern of the optimum vehicle state of the host vehicle The vehicle control device according to claim 4, wherein the vehicle is searched for.

Where v is the speed of the host vehicle at the time corresponding to the node, which is obtained based on the vehicle state represented by the node one hour before, and l is the speed of the host vehicle acquired by the vehicle information acquiring unit. This is the road speed limit obtained based on the position information.   6. Control means for controlling the driving of the host vehicle based on a vehicle state one hour later in the time series pattern of the optimal vehicle state of the host vehicle searched by the optimal solution searching unit. The vehicle control device according to claim 1. Computer
Vehicle information acquisition means for acquiring vehicle information including position information for each of the own vehicle and other vehicles, and each of nodes representing a plurality of types of vehicle states including an acceleration state, a deceleration state, a constant speed state, and a lane change state For each time up to a predetermined time, and in a vehicle state network representing each of the time series patterns of the vehicle state of the host vehicle by connecting the nodes in time series, The speed, acceleration, and relative position with respect to the other vehicle at the time corresponding to the node, obtained based on the vehicle state to be represented and the vehicle information of the own vehicle and the other vehicle acquired by the vehicle information acquisition unit An optimal solution searching means for searching for a time series pattern of an optimal vehicle state of the host vehicle according to a cost calculated based on
Program to function as.

Images