CN113790903A - Static obstacle avoidance test data evaluation method for low-speed intelligent networked automobile in closed scene - Google Patents
Static obstacle avoidance test data evaluation method for low-speed intelligent networked automobile in closed scene Download PDFInfo
- Publication number
- CN113790903A CN113790903A CN202111080258.3A CN202111080258A CN113790903A CN 113790903 A CN113790903 A CN 113790903A CN 202111080258 A CN202111080258 A CN 202111080258A CN 113790903 A CN113790903 A CN 113790903A
- Authority
- CN
- China
- Prior art keywords
- obstacle
- vehicle
- value
- obstacle avoidance
- standard
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 109
- 238000012360 testing method Methods 0.000 title claims abstract description 46
- 230000003068 static effect Effects 0.000 title claims abstract description 22
- 238000011157 data evaluation Methods 0.000 title abstract description 4
- 230000008569 process Effects 0.000 claims abstract description 81
- 238000011156 evaluation Methods 0.000 claims abstract description 17
- 238000005259 measurement Methods 0.000 claims abstract description 5
- 238000004088 simulation Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 8
- 238000001514 detection method Methods 0.000 claims description 4
- 230000004888 barrier function Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 230000007812 deficiency Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M17/00—Testing of vehicles
- G01M17/007—Wheeled or endless-tracked vehicles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a static obstacle avoidance test data evaluation method for a low-speed intelligent networked automobile in a closed scene. And combining vehicle state data and sensing measurement data in the obstacle avoidance process, comprehensively performing item evaluation on the data acquired in the test process, and then performing comprehensive scoring. The method is used for evaluating whether the vehicle is safe and reliable in the process of avoiding static obstacles, and can be used for comparing the performance of different vehicle types. The invention can improve the standardization degree of the test, improve the test efficiency and realize scientific evaluation on the test result.
Description
Technical Field
The invention relates to the technical field of intelligent networked automobile scene test data processing, in particular to a static obstacle avoidance test data evaluation method for a low-speed intelligent networked automobile closed scene.
Background
In recent years, the intelligent networking automobile technology has entered the stages of rapid technology evolution and accelerated enterprise layout. Enterprises from industries such as traditional automobile manufacturing, internet, smart device, communication equipment manufacturing and the like have developed a great deal of deep collaboration. The various related technologies gradually move to the mature stage.
The intelligent networked automobile is a new generation automobile for driving decision through sensing the environment, and is provided with a large amount of environment sensing, network communication, intelligent decision making and auxiliary driving equipment. At present, an intelligent driving automobile does not have a uniform test scene standard, test fields and demonstration areas meeting different requirements are built in all places, and different types of test tasks can be carried out. The chinese invention patent application CN202110141980 discloses an arrangement method of closed test scenes, and the chinese invention patent application CN202011035182 discloses a dynamic closed test system of an intelligent networked automobile. For different test scenes and test systems, the problems of test data acquisition, processing, evaluation and the like are involved. In view of the above, chinese patent CN2016102582237 discloses a method and an apparatus for evaluating an obstacle detection result of an unmanned vehicle, and chinese patent CN2016111768023 discloses a method and a system for predicting a state of an obstacle vehicle. The methods process the intelligent networking automobile test data of different layers.
The low-speed running intelligent networked automobile for testing in the closed road is different from a medium-high speed automobile for running on a public road, and the low-speed automobile is often used in special areas of parks, docks, factories and mines, schools and communities to realize specific functions of cargo transportation, patrol and cleaning. Compared with a medium-high speed intelligent networked automobile, the vehicle is easier to popularize, so that the test requirement is very urgent.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides the method for evaluating the static obstacle avoidance test data of the low-speed intelligent internet vehicle in the closed scene, which can realize the technical performance evaluation of the low-speed intelligent internet vehicle, promote the standardization of the test process, improve the test efficiency and enable the test conclusion to be more scientific, thereby finding the technical short board of the vehicle and improving the product performance.
The purpose of the invention is realized by the following technical scheme.
The method for evaluating the static obstacle avoidance test data of the low-speed intelligent networked automobile in the closed scene comprises the following steps of:
step 1, collecting a vehicle obstacle sectional area sensing data set Vs, an obstacle distance sensing data set Ds, a satellite positioning vehicle position data set Dv, a satellite positioning obstacle position data set Dm, a speed sensor data set Us, a level meter sensing data set Hs and a time stamp data set from a vehicle data center; calculating to obtain a standard average value Vra of the relative deviation of the cross-sectional area sensing of the obstacle, a standard average value Dra of the relative deviation of the distance sensing of the obstacle, a standard minimum value Drmin of the relative deviation of the distance sensing of the obstacle, a speed stability standard value Urs, a speed maximum value Urmax, a sway degree stability standard value Hrs and a tilt standard maximum value Hrmax; obtaining a cross-sectional area value Vm of the obstacle through measurement; importing the data into an evaluation system according to the time stamp sequence;
step 1.1, establishing a three-dimensional rectangular coordinate system by taking the center of a sensor as a coordinate origin O; randomly generating a convex polyhedron which does not contain a coordinate origin and has a deficiency of 0 as a simulated obstacle;
calculating the geometric center of the convex polyhedron, wherein the coordinate is G (a, b and c), the point G is connected with the point O, and the vector OG (a, b and c) represents the direction pointed by the obstacle sensor;
generating a parallel plane cluster, and expressing the parallel plane cluster as ax + by + cz + dk ═ 0 by a formula, wherein dk is a real number and represents different intercepts; changing the value of dk, calculating the sectional area of the plane cluster and the simulated obstacle, finding the maximum value Smax of the sectional area, and recording the dk as d;
according to the minimum resolving distance dx and dy of the obstacle sensor given by the manual in the x and y directions, carrying out grid division on ax + by + cz + d as 0, wherein the size of each grid is Sg as dx;
and keeping St equal to 0, and distinguishing each grid divided on the plane: the cross section area of the barrier contained in the grid is Sc, and if Sc/Sg > is 50%, St is St + 1;
the process is circulated, and all grids are traversed; calculating Sp (i) ═ Abs [ (St-Sc)/Sc ], wherein i represents the simulation of the number of times and Abs [ ] represents the absolute value;
a total of NN simulations were performed, and the standard deviation average Vra ═ sp (i) >, where < > denotes the calculated average, of the obstacle cross-sectional area sensing.
Step 1.2, generating a concentric spherical cluster by taking an original point O as a center, and expressing the concentric spherical cluster as x ^2+ y ^2 ^ rk ^2 with a formula, wherein rk is a real number greater than zero and expresses different radiuses;
changing the value of rk from small to large, and when there is an intersection point between rk and the surface of the simulated obstacle, marking the intersection point as C (cx, cy, cz), and marking the value of rk as r;
calculating [ nn, mm ] ═ div [ r, dz ], div [ ] representing an integer division operation, dd representing a quotient of the integer division, mm representing a remainder, according to a minimum resolving distance dz in the z direction of an obstacle sensor given in a manual, and dd +1 if mm > -dz/2;
altogether NN simulations were performed, dp (i) ═ Abs [ (dd-r)/r ], where i denotes the simulation for the number of times, the obstacle distance sensing relative deviation criterion mean value Dra ═ < dp (i) >, and the obstacle distance sensing relative deviation criterion minimum value Drmin ═ min [ dp (i) >, where min [ ] denotes the minimum value taken.
Step 1.3 measures a vehicle speed data set Vs through a real vehicle test platform, calculates a stability criterion value Urs ═ Std [ Vs ], calculates a speed maximum value Urmax [ Vs ], calculates a stability criterion value Hrs ═ Std [ Ts ], and calculates a tilt criterion maximum value Hrmax [ Ts ].
And 2, detecting and evaluating the cross section of the obstacle, calculating a standard average value A1 of the sensing relative deviation of the cross section of the obstacle, and judging whether A1 exceeds a specified standard.
According to the vehicle obstacle cross-sectional area sensing data set Vs, the measured obstacle cross-sectional area value Vm and the obstacle cross-sectional area sensing relative deviation standard average value Vra, the obstacle cross-sectional area sensing relative deviation standard average value A1 is < abs [ (Vs-Vm)/Vm ] >, abs [ ] represents an absolute value, and < > represents an average value.
And 3, detecting and evaluating the relative distance of the obstacle, calculating the average value A2 of the relative deviation of the obstacle distance sensing, and judging whether A2 exceeds a specified standard.
The method comprises the steps of obtaining a vehicle obstacle distance sensing data set Ds, a satellite positioning vehicle position data set Dv, a satellite positioning obstacle position data set Dm and an obstacle distance sensing relative deviation standard average value Dra, wherein the obstacle distance sensing relative deviation average value A2 ═ abs [ (Ds-Dd)/Dd ] >, Dd ═ dist [ Dv-Dm ], dist [ ] represents a distance, abs [ ] represents an absolute value, and < > represents an average value.
And 4, evaluating the vehicle speed stability in the obstacle avoidance process, calculating a vehicle speed deviation value A3 in the obstacle avoidance process, and judging whether the vehicle speed fluctuation amplitude exceeds a specified standard or not and whether the vehicle speed change is stable or not.
According to the vehicle speed sensor data set Us and the obstacle avoidance process vehicle speed stability standard value Urs, the obstacle avoidance process vehicle speed deviation value A3 is std [ Us ]/Urs, and std [ x ] represents the calculated relative standard deviation.
And 5, evaluating the vehicle swing stability in the obstacle avoidance process, calculating a vehicle swing degree deviation value A4 in the obstacle avoidance process, and judging whether the vehicle inclination swing degree exceeds a specified standard in the obstacle avoidance process.
According to the vehicle level meter sensing data set Hs and the vehicle swing degree stability standard value Hrs in the obstacle avoidance process, the vehicle swing degree deviation value A4 is std [ Hs ]/Hrs, and std [ x ] represents the calculated relative standard deviation.
And 6, evaluating the minimum distance in the obstacle avoidance process, calculating the minimum value A5 of the relative deviation of the obstacle distance sensing, and judging whether the vehicle is too close to the obstacle in the obstacle avoidance process.
And when A5 is smaller than Drmin, judging that the vehicle is unqualified when being too close to the obstacle in the obstacle avoidance process.
And 7, evaluating the maximum inclination degree of the vehicle in the obstacle avoidance process, calculating the maximum inclination degree A6 of the vehicle in the obstacle avoidance process, and judging whether the maximum inclination angle of the vehicle in the obstacle avoidance process exceeds a specified standard.
The method comprises the steps that according to a vehicle level meter sensing data set Hs and a maximum vehicle inclination standard value Hrmax in an obstacle avoidance process, the average value A6 of the relative deviation of the obstacle distance sensing is max [ Hs ], and when A6 is larger than Hrmax, the situation that the inclination degree of the vehicle is too large and unqualified in the obstacle avoidance process is judged.
And 8, evaluating the maximum vehicle speed in the obstacle avoidance process, calculating the maximum vehicle speed A7 in the obstacle avoidance process, and judging whether the maximum vehicle speed exceeds the standard or not.
According to the vehicle speed sensor data set Us and the vehicle speed maximum value Urmax, the vehicle maximum speed A7 in the obstacle avoidance process is max [ Us ], and when A7 is larger than Urmax, the fact that the vehicle speed exceeds the maximum value in the obstacle avoidance process is judged to be unqualified.
And 9, checking the values A1-A7, judging whether the vehicle is unqualified in the obstacle avoidance task, and outputting the judgment result to a paper or electronic storage medium.
Step 10, calculating four weighted values:
w1＝Vra/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w2＝Dra/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w3＝(Urs/<Us>)/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w4＝(Hrs/<Hs>)/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)；
and calculating a comprehensive evaluation result w-A1 w1+ A2 w2+ A3 w3+ A4 w4, wherein the values of w1, w2, w3 and w4 are all between 0 and 1, and outputting the evaluation result to a paper or electronic storage medium.
Wherein the vehicles are all low-speed intelligent networked automobiles; the obstacles are all static and simulated obstacles arranged in a test field, and comprise: one or more models or entities among a fault car, a hard shoulder, a well lid, a warning sign, a simulated dummy.
Compared with the prior art, the invention has the advantages that:
(1) a method for evaluating static obstacle avoidance test data of a low-speed intelligent networked automobile in a closed scene is provided.
(2) The scheme can promote the standardization degree of the data processing flow of the test task.
(3) The method is beneficial to improving the testing efficiency and enabling the testing conclusion to be more scientific, so that the technical short board of the vehicle is found and the product performance is improved.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The invention is described in detail below with reference to the drawings and specific examples.
As shown in fig. 1, the method for evaluating the static obstacle avoidance test data of the low-speed intelligent networked automobile in the closed scene comprises the following steps:
step 1, collecting a vehicle obstacle sectional area sensing data set Vs, an obstacle distance sensing data set Ds, a satellite positioning vehicle position data set Dv, a satellite positioning obstacle position data set Dm, a speed sensor data set Us, a level meter sensing data set Hs and a time stamp data set from a vehicle data center; calculating to obtain a standard average value Vra of the relative deviation of the cross-sectional area sensing of the obstacle, a standard average value Dra of the relative deviation of the distance sensing of the obstacle, a standard minimum value Drmin of the relative deviation of the distance sensing of the obstacle, a speed stability standard value Urs, a speed maximum value Urmax, a sway degree stability standard value Hrs and a tilt standard maximum value Hrmax; obtaining a cross-sectional area value Vm of the obstacle through measurement; importing the data into an evaluation system according to the time stamp sequence;
step 1.1, establishing a three-dimensional rectangular coordinate system by taking the center of a sensor as a coordinate origin O; randomly generating a convex polyhedron which does not contain a coordinate origin and has a deficiency of 0 as a simulated obstacle;
calculating the geometric center of the convex polyhedron, wherein the coordinate is G (a, b and c), the point G is connected with the point O, and the vector OG (a, b and c) represents the direction pointed by the obstacle sensor;
generating a parallel plane cluster, and expressing the parallel plane cluster as ax + by + cz + dk ═ 0 by a formula, wherein dk is a real number and represents different intercepts; changing the value of dk, calculating the sectional area of the plane cluster and the simulated obstacle, finding the maximum value Smax of the sectional area, and recording the dk as d;
according to the minimum resolving distance dx and dy of the obstacle sensor given by the manual in the x and y directions, carrying out grid division on ax + by + cz + d as 0, wherein the size of each grid is Sg as dx;
and keeping St equal to 0, and distinguishing each grid divided on the plane: the cross section area of the barrier contained in the grid is Sc, and if Sc/Sg > is 50%, St is St + 1;
the process is circulated, and all grids are traversed; calculating Sp (i) ═ Abs [ (St-Sc)/Sc ], wherein i represents the simulation of the number of times and Abs [ ] represents the absolute value;
a total of NN simulations were performed, and the standard deviation average Vra ═ sp (i) >, where < > denotes the calculated average, of the obstacle cross-sectional area sensing.
Step 1.2, generating a concentric spherical cluster by taking an original point O as a center, and expressing the concentric spherical cluster as x ^2+ y ^2 ^ rk ^2 with a formula, wherein rk is a real number greater than zero and expresses different radiuses;
changing the value of rk from small to large, and when there is an intersection point between rk and the surface of the simulated obstacle, marking the intersection point as C (cx, cy, cz), and marking the value of rk as r;
calculating [ nn, mm ] ═ div [ r, dz ], div [ ] representing an integer division operation, dd representing a quotient of the integer division, mm representing a remainder, according to a minimum resolving distance dz in the z direction of an obstacle sensor given in a manual, and dd +1 if mm > -dz/2;
altogether NN simulations were performed, dp (i) ═ Abs [ (dd-r)/r ], where i denotes the simulation for the number of times, the obstacle distance sensing relative deviation criterion mean value Dra ═ < dp (i) >, and the obstacle distance sensing relative deviation criterion minimum value Drmin ═ min [ dp (i) >, where min [ ] denotes the minimum value taken.
Step 1.3 measures a vehicle speed data set Vs through a real vehicle test platform, calculates a stability criterion value Urs ═ Std [ Vs ], calculates a speed maximum value Urmax [ Vs ], calculates a stability criterion value Hrs ═ Std [ Ts ], and calculates a tilt criterion maximum value Hrmax [ Ts ].
And 2, detecting and evaluating the cross section of the obstacle, calculating a standard average value A1 of the sensing relative deviation of the cross section of the obstacle, and judging whether A1 exceeds a specified standard.
According to the vehicle obstacle cross-sectional area sensing data set Vs, the measured obstacle cross-sectional area value Vm and the obstacle cross-sectional area sensing relative deviation standard average value Vra, the obstacle cross-sectional area sensing relative deviation standard average value A1 is < abs [ (Vs-Vm)/Vm ] >, abs [ ] represents an absolute value, and < > represents an average value.
And 3, detecting and evaluating the relative distance of the obstacle, calculating the average value A2 of the relative deviation of the obstacle distance sensing, and judging whether A2 exceeds a specified standard.
The method comprises the steps of obtaining a vehicle obstacle distance sensing data set Ds, a satellite positioning vehicle position data set Dv, a satellite positioning obstacle position data set Dm and an obstacle distance sensing relative deviation standard average value Dra, wherein the obstacle distance sensing relative deviation average value A2 ═ abs [ (Ds-Dd)/Dd ] >, Dd ═ dist [ Dv-Dm ], dist [ ] represents a distance, abs [ ] represents an absolute value, and < > represents an average value.
And 4, evaluating the vehicle speed stability in the obstacle avoidance process, calculating a vehicle speed deviation value A3 in the obstacle avoidance process, and judging whether the vehicle speed fluctuation amplitude exceeds a specified standard or not and whether the vehicle speed change is stable or not.
According to the vehicle speed sensor data set Us and the obstacle avoidance process vehicle speed stability standard value Urs, the obstacle avoidance process vehicle speed deviation value A3 is std [ Us ]/Urs, and std [ x ] represents the calculated relative standard deviation.
And 5, evaluating the vehicle swing stability in the obstacle avoidance process, calculating a vehicle swing degree deviation value A4 in the obstacle avoidance process, and judging whether the vehicle inclination swing degree exceeds a specified standard in the obstacle avoidance process.
According to the vehicle level meter sensing data set Hs and the vehicle swing degree stability standard value Hrs in the obstacle avoidance process, the vehicle swing degree deviation value A4 is std [ Hs ]/Hrs, and std [ x ] represents the calculated relative standard deviation.
And 6, evaluating the minimum distance in the obstacle avoidance process, calculating the minimum value A5 of the relative deviation of the obstacle distance sensing, and judging whether the vehicle is too close to the obstacle in the obstacle avoidance process.
And when A5 is smaller than Drmin, judging that the vehicle is unqualified when being too close to the obstacle in the obstacle avoidance process.
And 7, evaluating the maximum inclination degree of the vehicle in the obstacle avoidance process, calculating the maximum inclination degree A6 of the vehicle in the obstacle avoidance process, and judging whether the maximum inclination angle of the vehicle in the obstacle avoidance process exceeds a specified standard.
The method comprises the steps that according to a vehicle level meter sensing data set Hs and a maximum vehicle inclination standard value Hrmax in an obstacle avoidance process, the average value A6 of the relative deviation of the obstacle distance sensing is max [ Hs ], and when A6 is larger than Hrmax, the situation that the inclination degree of the vehicle is too large and unqualified in the obstacle avoidance process is judged.
And 8, evaluating the maximum vehicle speed in the obstacle avoidance process, calculating the maximum vehicle speed A7 in the obstacle avoidance process, and judging whether the maximum vehicle speed exceeds the standard or not.
According to the vehicle speed sensor data set Us and the vehicle speed maximum value Urmax, the vehicle maximum speed A7 in the obstacle avoidance process is max [ Us ], and when A7 is larger than Urmax, the fact that the vehicle speed exceeds the maximum value in the obstacle avoidance process is judged to be unqualified.
And 9, checking the values A1-A7, judging whether the vehicle is unqualified in the obstacle avoidance task, and outputting the judgment result to a paper or electronic storage medium.
Step 10, calculating four weighted values:
w1＝Vra/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w2＝Dra/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w3＝(Urs/<Us>)/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w4＝(Hrs/<Hs>)/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)；
and calculating a comprehensive evaluation result w-A1 w1+ A2 w2+ A3 w3+ A4 w4, wherein the values of w1, w2, w3 and w4 are all between 0 and 1, and outputting the evaluation result to a paper or electronic storage medium.
Wherein the vehicles are all low-speed intelligent networked automobiles; the obstacles are all static and simulated obstacles arranged in a test field, and comprise: one or more models or entities among a fault car, a hard shoulder, a well lid, a warning sign, a simulated dummy.
Example 1
Step 1, collecting a vehicle obstacle sectional area sensing data set Vs, an obstacle distance sensing data set Ds, a satellite positioning vehicle position data set Dv, a satellite positioning obstacle position data set Dm, a speed sensor data set Us, a level meter sensing data set Hs and a time stamp data set from a vehicle data center; searching from a performance requirement manual to obtain a standard average value Vra of the sensing relative deviation of the cross-sectional area of the obstacle, which is 0.05 square meter, a standard average value Dra of the sensing relative deviation of the distance of the obstacle, which is 0.03, a standard minimum value Drmin of the sensing relative deviation of the distance of the obstacle, which is 0.2 meter, a speed stability standard value Urs, which is 1 kilometer per hour, a speed maximum value Urmax, which is 30 kilometers per hour, a sway degree stability standard value Hrs, which is 0.1 degree, and a tilt standard maximum value Hrmax, which is 3 degrees; obtaining a cross-sectional area value Vm of the obstacle as 0.31 square meter through measurement; and importing the data into the evaluation system according to the time stamp sequence.
And 2, detecting and evaluating the cross section of the obstacle, and calculating the standard average value A1 of the relative deviation of the cross section sensing of the obstacle to be 0.08.
And 3, detecting and evaluating the relative distance of the obstacle, and calculating the average value A2 of the relative deviation of the obstacle distance sensing to be 0.05.
And 4, evaluating the stability of the vehicle speed in the obstacle avoidance process, and calculating a vehicle speed deviation value A3 of 1.39 kilometers per hour in the obstacle avoidance process.
And 5, evaluating the vehicle swing stability in the obstacle avoidance process, and calculating a vehicle swing degree deviation value A4 to be 0.15 degrees in the obstacle avoidance process.
And 6, evaluating the minimum distance in the obstacle avoidance process, calculating the minimum value A5 of the relative deviation of the obstacle distance sensing, which is 0.3 m and is larger than the value of Drmin, and judging that the vehicle is not too close to the obstacle in the obstacle avoidance process.
And 7, evaluating the maximum inclination degree of the vehicle in the obstacle avoidance process, calculating the maximum inclination degree A6 of the vehicle in the obstacle avoidance process to be 0.4 degrees and less than Hrmax, and judging that the maximum inclination angle of the vehicle in the obstacle avoidance process does not exceed a specified standard.
And 8, evaluating the maximum vehicle speed in the obstacle avoidance process, calculating the maximum vehicle speed A7 in the obstacle avoidance process to be 25.3 km/h and less than the value of Urmax, and judging that the maximum vehicle speed does not exceed the standard.
And 9, checking the values of A1-A7, judging that the vehicle does not have the condition that the detection items are not qualified in the obstacle avoidance task, and storing the judgment result into a paper or electronic storage medium.
And 10, calculating the values of w1, w2, w3 and w4 to be 0.25, calculating the comprehensive evaluation result w-A1 w1+ A2 w2+ A3 w3+ A4 w4 to be 0.39, and outputting the evaluation result to the paper or electronic storage medium.
Claims (10)
1. The method for evaluating the static obstacle avoidance test data of the low-speed intelligent networked automobile in the closed scene comprises the following steps of:
1) collecting a vehicle obstacle sectional area sensing data set Vs, an obstacle distance sensing data set Ds, a satellite positioning vehicle position data set Dv, a satellite positioning obstacle position data set Dm, a speed sensor data set Us, a level sensing data set Hs and a time stamp data set from a vehicle data center; calculating to obtain a standard average value Vra of the relative deviation of the cross-sectional area sensing of the obstacle, a standard average value Dra of the relative deviation of the distance sensing of the obstacle, a standard minimum value Drmin of the relative deviation of the distance sensing of the obstacle, a speed stability standard value Urs, a speed maximum value Urmax, a sway degree stability standard value Hrs and a tilt standard maximum value Hrmax; obtaining a cross-sectional area value Vm of the obstacle through measurement; importing the data into an evaluation system according to the time stamp sequence;
2) and (3) detecting and evaluating the cross section of the obstacle: calculating a standard average value A1 of the sensing relative deviation of the cross section area of the obstacle, and judging whether A1 exceeds a specified standard;
3) obstacle relative distance detection evaluation: calculating the average value A2 of the relative deviation of the obstacle distance sensing, and judging whether A2 exceeds a specified standard;
4) evaluating the vehicle speed stability in the obstacle avoidance process, calculating a vehicle speed deviation value A3 in the obstacle avoidance process, and judging whether the vehicle speed fluctuation amplitude exceeds a specified standard and whether the vehicle speed change is stable;
5) evaluating the vehicle swing stability in the obstacle avoidance process, calculating a vehicle swing degree deviation value A4 in the obstacle avoidance process, and judging whether the vehicle inclination swing degree exceeds a specified standard in the obstacle avoidance process;
6) evaluating the minimum distance in the obstacle avoidance process, calculating the minimum value A5 of the relative deviation of the obstacle distance sensing, and judging whether the vehicle is too close to the obstacle in the obstacle avoidance process;
7) evaluating the maximum inclination degree of the vehicle in the obstacle avoidance process, calculating the maximum inclination degree A6 of the vehicle in the obstacle avoidance process, and judging whether the maximum inclination angle of the vehicle in the obstacle avoidance process exceeds a specified standard;
8) evaluating the maximum vehicle speed in the obstacle avoidance process, calculating the maximum vehicle speed A7 in the obstacle avoidance process, and judging whether the maximum vehicle speed exceeds the standard or not;
9) checking the values of A1-A7, judging whether the vehicle is unqualified in the obstacle avoidance task, and outputting the judgment result to a paper or electronic storage medium;
10) calculating a comprehensive evaluation result w-A1 w1+ A2 w2+ A3 w3+ A4 w4, wherein w1+ w2+ w3+ w4 is 1, and outputting the evaluation result to a paper or electronic storage medium;
wherein the vehicles are all low-speed intelligent networked automobiles; the obstacles are all static and simulated obstacles arranged in a test field, and comprise: one or more models or entities among a fault car, a hard shoulder, a well lid, a warning sign, a simulated dummy.
2. The method for evaluating the static obstacle avoidance test data of the low-speed intelligent networked automobile closed scene according to claim 1, wherein the step 1) is specifically as follows:
1.1) establishing a three-dimensional rectangular coordinate system by taking the center of the sensor as a coordinate origin O; randomly generating a convex polyhedron which does not contain a coordinate origin and has a deficiency of 0 as a simulated obstacle;
calculating the geometric center of the convex polyhedron, wherein the coordinate is G (a, b and c), the point G is connected with the point O, and the vector OG (a, b and c) represents the direction pointed by the obstacle sensor;
generating a parallel plane cluster, and expressing the parallel plane cluster as ax + by + cz + dk ═ 0 by a formula, wherein dk is a real number and represents different intercepts; changing the value of dk, calculating the sectional area of the plane cluster and the simulated obstacle, finding the maximum value Smax of the sectional area, and recording the dk as d;
according to the minimum resolving distance dx and dy of the obstacle sensor given by the manual in the x and y directions, carrying out grid division on ax + by + cz + d as 0, wherein the size of each grid is Sg as dx;
and keeping St equal to 0, and distinguishing each grid divided on the plane: the cross section area of the barrier contained in the grid is Sc, and if Sc/Sg > is 50%, St is St + 1;
the process is circulated, and all grids are traversed; calculating Sp (i) ═ Abs [ (St-Sc)/Sc ], wherein i represents the simulation of the number of times and Abs [ ] represents the absolute value;
performing NN simulations in total, and calculating a standard deviation average value Vra ═ Sp (i) >, wherein the Vra represents the calculated average value;
1.2) generating a concentric spherical cluster by taking an original point O as a center, wherein the concentric spherical cluster is expressed as x ^2+ y ^2 ^ rk ^2, wherein rk is a real number greater than zero and represents different radiuses;
changing the value of rk from small to large, and when there is an intersection point between rk and the surface of the simulated obstacle, marking the intersection point as C (cx, cy, cz), and marking the value of rk as r;
calculating [ nn, mm ] ═ div [ r, dz ], div [ ] representing an integer division operation, dd representing a quotient of the integer division, mm representing a remainder, according to a minimum resolving distance dz in the z direction of an obstacle sensor given in a manual, and dd +1 if mm > -dz/2;
performing NN simulations in total, calculating Dp (i) ═ Abs [ (dd-r)/r ], wherein i represents the simulation of the second time, calculating an obstacle distance sensing relative deviation standard average value Dra ═ Dp (i) >, and calculating an obstacle distance sensing relative deviation standard minimum value Drmin ═ min [ Dp (i) >, wherein min [ ] represents taking the minimum value;
1.3) measuring a vehicle speed data set Vs through a real vehicle testing platform, calculating a stability standard value Urs which is Std [ Vs ], calculating a speed maximum value Urmax [ Vs ], calculating a stability standard value Hrs which is Std [ Ts ] and a tilt standard maximum value Hrmax which is max [ Ts ].
3. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile according to claim 2, wherein according to the vehicle obstacle sectional area sensing data set Vs, the measured obstacle sectional area value Vm and the standard average value of the obstacle sectional area sensing relative deviation Vra, the standard average value of the obstacle sectional area sensing relative deviation A1 is < abs [ (Vs-Vm)/Vm ] >, abs [ ] represents an absolute value, and < > represents an average value; and when the A1 is larger than the Vra, judging that the cross-sectional area of the vehicle obstacle is detected ineligibly.
4. The method for evaluating static obstacle avoidance test data of a closed scene of a low-speed intelligent networked automobile according to claim 2, wherein according to the vehicle obstacle distance sensing data set Ds, the satellite-positioned vehicle position data set Dv, the satellite-positioned obstacle position data set Dm, and the obstacle distance sensing relative deviation standard average value Dra, the obstacle distance sensing relative deviation average value a2 ═ abs [ (Ds-Dd)/Dd ] >, Dd ═ dist [ Dv-Dm ], dist [ ] represents a distance calculation, abs [ ] represents an absolute value, and < > represents an average value; and when the A2 is larger than the Dra, judging that the vehicle is unqualified for the obstacle distance detection.
5. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile according to claim 2, wherein according to the vehicle speed sensor data set Us and the obstacle avoidance process vehicle speed stability standard value Urs, the obstacle avoidance process vehicle speed deviation value A3 ═ std [ Us ], std [ ] represents a calculation standard deviation, and when A3 is greater than Urs, the vehicle is judged to have unqualified speed fluctuation amplitude in the obstacle avoidance process.
6. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile according to claim 2, wherein according to the vehicle level meter sensing data set Hs and the obstacle avoidance process vehicle sway degree stability standard value Hrs, the vehicle sway degree deviation value A4 ═ std [ Hs ], std [ ] represents a calculation standard deviation, and when A4 is greater than Hrs, the vehicle sway degree is judged to be too large and unqualified in the obstacle avoidance process.
7. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile according to claim 2, wherein according to the vehicle obstacle distance sensing data set Ds, the obstacle distance sensing relative deviation standard minimum value Drmin, and the obstacle distance sensing relative deviation minimum value A5 ═ min [ Ds ], when A5 is smaller than Drmin, it is judged that the vehicle is unqualified to be too close to the obstacle in the obstacle avoidance process.
8. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile as claimed in claim 2, wherein the vehicle is judged to be unqualified due to the fact that the inclination degree of the vehicle is too large in the obstacle avoidance process when A6 is larger than Hrmax according to the vehicle level meter sensing data set Hs and the maximum vehicle inclination standard value Hrmax in the obstacle avoidance process, and the average value of the relative obstacle distance sensing deviation A6 is max [ Hs ].
9. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile according to claim 2, wherein according to the vehicle speed sensor data set Us and the vehicle speed maximum value Urmax, the vehicle maximum speed A7 ═ max [ Us ] in the obstacle avoidance process is judged, and when A7 is larger than Urmax, the vehicle speed exceeding the maximum value in the obstacle avoidance process is judged to be unqualified.
10. The method for evaluating the static obstacle avoidance test data of the closed scene of the low-speed intelligent networked automobile according to any one of claims 1 to 9, wherein four weighted values are calculated:
w1＝Vra/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w2＝Dra/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w3＝(Urs/<Us>)/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)，
w4＝(Hrs/<Hs>)/(Vra+Dra+Urs/<Us>+Hrs/<Hs>)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111080258.3A CN113790903A (en) | 2021-09-15 | 2021-09-15 | Static obstacle avoidance test data evaluation method for low-speed intelligent networked automobile in closed scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111080258.3A CN113790903A (en) | 2021-09-15 | 2021-09-15 | Static obstacle avoidance test data evaluation method for low-speed intelligent networked automobile in closed scene |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113790903A true CN113790903A (en) | 2021-12-14 |
Family
ID=79183488
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111080258.3A Pending CN113790903A (en) | 2021-09-15 | 2021-09-15 | Static obstacle avoidance test data evaluation method for low-speed intelligent networked automobile in closed scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113790903A (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106525454A (en) * | 2016-11-18 | 2017-03-22 | 特路（北京）科技有限公司 | Static obstacle response performance test method of automatic driving vehicle and test field |
| CN107103104A (en) * | 2016-02-19 | 2017-08-29 | 苏州溥诺斯智能科技有限公司 | A kind of Vehicular intelligent net translocation test system based on cross-layer cooperated framework |
| CN109598066A (en) * | 2018-12-05 | 2019-04-09 | 百度在线网络技术（北京）有限公司 | Effect evaluation method, device, equipment and the storage medium of prediction module |
| CN109827781A (en) * | 2018-12-29 | 2019-05-31 | 中国汽车技术研究中心有限公司 | A kind of vehicle testing data processing method, apparatus and system |
| CN110287832A (en) * | 2019-06-13 | 2019-09-27 | 北京百度网讯科技有限公司 | High-Speed Automatic Driving Scene barrier perception evaluating method and device |
| CN112195695A (en) * | 2020-09-27 | 2021-01-08 | 四川紫荆花开智能网联汽车科技有限公司 | Intelligent network-connected automobile dynamic closed test system |
| CN112466123A (en) * | 2021-02-02 | 2021-03-09 | 四川紫荆花开智能网联汽车科技有限公司 | Method for arranging intelligent networking automobile test scene in closed test field |
| CN112525542A (en) * | 2020-10-26 | 2021-03-19 | 泉州装备制造研究所 | Intelligent vehicle performance test system and method thereof |
| CN112781887A (en) * | 2020-12-21 | 2021-05-11 | 苏州挚途科技有限公司 | Method, device and system for testing vehicle performance |
| CN113022702A (en) * | 2021-04-29 | 2021-06-25 | 吉林大学 | Intelligent networking automobile self-adaptive obstacle avoidance system based on steer-by-wire and game result |
| KR20210088457A (en) * | 2020-12-10 | 2021-07-14 | 베이징 바이두 넷컴 사이언스 테크놀로지 컴퍼니 리미티드 | Road test method and apparatus for autonomous driving vehicle, device and storage medium |
| CN113359673A (en) * | 2020-06-29 | 2021-09-07 | 史秋虹 | Automatic driving automobile performance judgment system based on big data |
-
2021
- 2021-09-15 CN CN202111080258.3A patent/CN113790903A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107103104A (en) * | 2016-02-19 | 2017-08-29 | 苏州溥诺斯智能科技有限公司 | A kind of Vehicular intelligent net translocation test system based on cross-layer cooperated framework |
| CN106525454A (en) * | 2016-11-18 | 2017-03-22 | 特路（北京）科技有限公司 | Static obstacle response performance test method of automatic driving vehicle and test field |
| CN109598066A (en) * | 2018-12-05 | 2019-04-09 | 百度在线网络技术（北京）有限公司 | Effect evaluation method, device, equipment and the storage medium of prediction module |
| CN109827781A (en) * | 2018-12-29 | 2019-05-31 | 中国汽车技术研究中心有限公司 | A kind of vehicle testing data processing method, apparatus and system |
| CN110287832A (en) * | 2019-06-13 | 2019-09-27 | 北京百度网讯科技有限公司 | High-Speed Automatic Driving Scene barrier perception evaluating method and device |
| CN113359673A (en) * | 2020-06-29 | 2021-09-07 | 史秋虹 | Automatic driving automobile performance judgment system based on big data |
| CN112195695A (en) * | 2020-09-27 | 2021-01-08 | 四川紫荆花开智能网联汽车科技有限公司 | Intelligent network-connected automobile dynamic closed test system |
| CN112525542A (en) * | 2020-10-26 | 2021-03-19 | 泉州装备制造研究所 | Intelligent vehicle performance test system and method thereof |
| KR20210088457A (en) * | 2020-12-10 | 2021-07-14 | 베이징 바이두 넷컴 사이언스 테크놀로지 컴퍼니 리미티드 | Road test method and apparatus for autonomous driving vehicle, device and storage medium |
| CN112781887A (en) * | 2020-12-21 | 2021-05-11 | 苏州挚途科技有限公司 | Method, device and system for testing vehicle performance |
| CN112466123A (en) * | 2021-02-02 | 2021-03-09 | 四川紫荆花开智能网联汽车科技有限公司 | Method for arranging intelligent networking automobile test scene in closed test field |
| CN113022702A (en) * | 2021-04-29 | 2021-06-25 | 吉林大学 | Intelligent networking automobile self-adaptive obstacle avoidance system based on steer-by-wire and game result |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dhiman et al. | Pothole detection using computer vision and learning | |
| CN110689723B (en) | Truck overload identification method based on power distribution and self-learning | |
| CN110032949A (en) | A kind of target detection and localization method based on lightweight convolutional neural networks | |
| CN107316048A (en) | Point cloud classifications method and device | |
| CN112883820B (en) | Road target 3D detection method and system based on laser radar point cloud | |
| CN106446834A (en) | Vehicle type identification method and vehicle type identification device based on images | |
| CN114170448A (en) | Evaluation method and device for visual perception algorithm | |
| CN112578673B (en) | Perception decision and tracking control method for multi-sensor fusion of formula-free racing car | |
| Dabbiru et al. | Lidar data segmentation in off-road environment using convolutional neural networks (cnn) | |
| CN109885046B (en) | Mobile robot positioning and accelerating method based on particle filtering | |
| CN113221682A (en) | Bridge vehicle load space-time distribution fine-grained identification method based on computer vision | |
| CN111964763B (en) | Method for detecting intermittent driving behavior of automobile in weighing area of dynamic flat-plate scale | |
| CN114299011A (en) | Remote sensing target quadrilateral frame rapid detection method based on deep learning | |
| Camilleri et al. | Detecting road potholes using computer vision techniques | |
| CN112800934A (en) | Behavior identification method and device for multi-class engineering vehicle | |
| Shih et al. | Reconstruction and synthesis of lidar point clouds of spray | |
| CN113790903A (en) | Static obstacle avoidance test data evaluation method for low-speed intelligent networked automobile in closed scene | |
| CN115964615A (en) | Track smoothness evaluation method based on centerline point cloud data | |
| CN113447953B (en) | Background filtering method based on road traffic point cloud data | |
| CN113392695B (en) | Highway truck and wheel axle identification method thereof | |
| CN111091565B (en) | Self-adaptive motion characteristic matching and recognition bow net contact point detection method | |
| CN114004740A (en) | Building wall line extraction method based on unmanned aerial vehicle laser radar point cloud | |
| CN112651986A (en) | Environment recognition method, recognition device, recognition system, electronic device, and medium | |
| Altmann et al. | Identification of Damaging Road Events Using Pseudo Damage Density | |
| Lin et al. | Traffic Object Detection in Virtual Environments |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |