Autonomous travel system based on artificial intelligence and automatic driving technology

Abstract

The invention relates to the technical field of intelligent travel, in particular to an autonomous travel system based on artificial intelligence and an automatic driving technology, which is based on a WeChat applet, takes car-pooling-based automatic driving shared travel platform software as a carrier, comprehensively calculates a car calling request of a user and plans an automatic driving minibus scheduling route; the platform software has the functions of site selection, path planning, cost calculation, order adjustment and user authentication, and is tested by the WeChat development platform; the platform identifies various facilities based on an inverse address resolution and address resolution function, and provides rapid guidance for automatic driving bus scheduling in combination with a map; according to the invention, actual measurement is carried out through the campus high-pass autonomous travel system, the driving roads in the campus are 100% covered, and the feasibility of the high-pass autonomous travel system is verified; the system and the original fixed station connection in the campus are mutually supplemented, convenient campus travel reservation service can be provided, and practical application of'travel-as-service 'in a campus environment is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent travel, and particularly to an autonomous travel system based on artificial intelligence and autonomous driving technology. Background Art

[0002] At present, with the acceleration of urbanization and population growth, the problem of urban traffic congestion has become increasingly prominent, and traditional transportation methods can no longer meet people's travel needs. Future communities are the basic units for the high-quality development of future cities, with the basic connotations of green intensiveness, wisdom sharing, representing the ideal community life model for humanity, and are still in the exploration and experimentation stage. In recent years, places such as Hangzhou, Zhejiang Province have created future communities by building digital infrastructure, putting services on the cloud, and enabling unmanned delivery vehicles. However, as an important part of residents' daily lives, the concept of intelligent travel in future communities has not been fully designed.

[0003] At present, it is necessary to vigorously develop shared transportation and build a service system based on mobile intelligent terminal technology. Priority should be given to the development of human-machine shared driving and vehicle-road-cloud collaborative technologies for intelligent transportation systems. In the industrial field, companies such as Baidu, Huawei, Pony.ai, and Waymo have successively launched autonomous vehicles in Beijing, Guangzhou, Phoenix and other places to meet travel needs such as community shuttles, daily commuting, and airport pick-ups. With the development of technologies such as connected autonomous driving, cloud computing, and digital twin, the shared travel system shows a trend of high integration of passengers, vehicles, roads, and the cloud, and then deep integration with communities.

[0004] There have been many inventions of autonomous travel systems, but most of them focus on the vehicle body perception system, intelligent networking between vehicles, and the overall concept of the shuttle system. However, the solutions for passengers' travel needs are not perfect enough, and the technical solutions for realizing the integration of vehicle-road-network are not fully explored. The "human-vehicle-road-network" is not studied as an integrated shuttle system. Summary of the Invention

[0005] In order to make up for the deficiencies of the existing technology, the present invention proposes a technology for constructing a community-based high-access autonomous travel system. Combining users' personalized travel preferences, with a safety management and application service integrated cloud control platform as the core, using 5G+L4-level electric autonomous driving minibuses as the carrier, supplemented by intelligent roadside devices as the support, fully considering users' needs, and combining digital twin and simulation-based actual measurement optimization to construct a community-based high-access autonomous travel system, so as to realize the efficient operation of the "human-vehicle-road-network" four-in-one collaborative autonomous driving short-distance shuttle vehicle.

[0006] An autonomous travel system based on artificial intelligence and autonomous driving technology described in the present invention. This travel system is based on a WeChat mini-program and uses a carpool-based autonomous driving shared travel platform software as a carrier to comprehensively calculate user car-hailing requests and plan the scheduling route of autonomous driving minibuses. The platform software includes functions such as location selection, route planning, fare calculation, order adjustment, and user authentication, and has been tested on the WeChat development platform. The platform is based on reverse geocoding and geocoding functions to identify various facilities, and combines with a map to provide rapid guidance for the scheduling of autonomous driving minibuses. The platform provides a location search function based on a keyword search function. Users can switch the regional list according to their needs, freely adjust the search range, and quickly select a location by supplementing with historical records and keywords. After the location is selected, the platform obtains the coordinates of the shortest path point string in the driving mode based on the online map API and displays the route according to the map zoom factor. When used in conjunction with an autonomous driving minibus, the platform assists in scheduling based on the head direction, speed, and acceleration data in the vehicle package data to achieve optimal route planning.

[0007] Preferably, the platform calculates the ride fare based on a starting price + mileage and supports subsequent adjustment of the calculation method according to the operating costs of autonomous driving minibuses. After the order is confirmed, the platform will display the estimated arrival time of the vehicle, and users can freely adjust or cancel the order during the waiting period. To provide better travel services, the platform also provides a user authentication function. After the user logs in, they can obtain historical order records and perform personalized settings. User request data is also extracted from the cloud background.

[0008] Preferably, the system further includes an ultra-long-range physical twin remote control module. By measuring the 5G signal strength on the main road, a 5G signal visualization heat map is drawn, and coordinated with the operator to add base stations and strengthen the signal to provide a stable 5G communication environment, which serves as the digital base for the physical twin.

[0009] Preferably, during operation, in response to complex road traffic conditions, the on-vehicle control system of the autonomous driving minibus predicts the trajectories of surrounding moving targets based on multi-modal perception data, and judges whether the current traffic situation can be resolved autonomously by the minibus through the system. When the autonomous driving minibus cannot resolve the encountered emergencies autonomously, the on-vehicle system will use 5G high-speed communication to transmit remote videos and images to the cloud control platform and display them in the ultra-long-range physical twin system in the laboratory. The operator in the laboratory uses the ultra-long-range physical twin system for real-time operation by combining the instant remote images with the status of multi-modal on-vehicle perception data.

[0010] Preferably, the driver uses visible light and near-infrared to establish a common perception technology that adapts to the limb postures, head orientations, and visual attention of drivers under different lighting, operation, and traffic environments. Combining millimeter-wave radar and wearable devices, the fusion perception of the electromyogram, heart rate, and respiratory physiological characteristic information of the driver is realized.

[0011] The beneficial effects of the present invention are as follows:

[0012] 1. Through the actual measurement of the campus high-throughput autonomous travel system of the present invention, the driving roads within the campus are covered by 100%, verifying the feasibility of the high-throughput autonomous travel system; it complements the original fixed stations within the campus, can provide convenient campus travel reservation services, realizes the practical application of "travel as a service" in the campus environment, the call volume of the travel platform software is 10,000 times / day, the concurrency is 5, and the waiting time of users is actually reduced; it meets the personalized travel needs of users without fixed stations, fixed routes, and fixed timetables; the system encourages low-carbon travel modes, implements the principles of low-carbon emission reduction, energy conservation and high efficiency, and strongly promotes the sustainable development of campus life; in the future, the system is expected to develop into an autonomous travel service system with strong autonomy, high flexibility, and multi-modal characteristics.

[0013] 2. The present invention combines with the information service cloud control platform to construct a community-based high-throughput autonomous travel system with functions such as one-stop travel, reservation-based travel, autonomous service, instant demand response, and dynamic supply-demand matching; this system fully considers the travel needs of passengers, takes the integrated cloud control platform of safety management and application services as the core, uses 5G+L4-level electric autonomous driving minibuses as the carrier, uses high-precision maps as the digital base, and is supplemented by intelligent roadside devices as the support to realize the operation of a four-in-one collaborative autonomous driving short-distance shuttle bus of "people-vehicle-road-network".

[0014] 3. The present invention develops an autonomous driving shared travel platform software based on WeChat mini-programs and mainly for carpooling, comprehensively calculates users' car-hailing requests, and plans the dispatching routes of autonomous driving minibuses, aiming to provide low-carbon and convenient travel services; by collecting and processing laser point cloud data, drawing high-precision maps and conducting log inspections, it serves as the twin base for constructing a three-dimensional visual digital road space; uses intelligent roadside perception devices and vehicle-road collaborative facilities to build a smart traffic management platform, and finally realizes holographic intersections and digital twins remotely; uses high-precision maps as the basis of location information, obtains multi-modal perception data through in-vehicle devices, and gets instant environmental information; based on the prediction of the trajectories of autonomous driving vehicles and surrounding targets, combined with instant map information, realizes the collaborative perception of autonomous driving minibuses; when the unmanned minibus cannot solve emergencies independently, uses the cloud control platform for remote video transmission and displays it in the ultra-long-range physical twin system; combined with the state of multi-modal in-vehicle perception data, uses the ultra-long-range physical twin system for remote real-time operation to realize the temporary takeover and remote control of the minibus. Description of the Drawings

[0015] The present invention will be further described below in conjunction with the drawings and embodiments.

[0016] Figure 1 It is the architecture diagram of the community-based high-throughput autonomous travel system in the present invention.

[0017] Figure 2 It is the "autonomous driving minibus sharing travel platform" module in the present invention.

[0018] Figure 3 It is the "high-precision map drawing, intelligent roadside device perception and visualization" module in the present invention.

[0019] Figure 4 It is the "autonomous driving minibus operation environment perception and trajectory prediction" module in the present invention.

[0020] Figure 5 It is the "beyond visual range physical twin remote control" module in the present invention. Detailed implementation manners

[0021] In order to make the technical means, creative features, achieved purposes and functions realized by the present invention easy to understand, the present invention will be further described below in conjunction with specific implementation manners.

[0022] As Figures 1-5 shown, an autonomous travel system based on artificial intelligence and autonomous driving technology described in the present invention is specifically described as follows:

[0023] The present invention meets the normal operation requirements through intelligent connected autonomous driving minibuses at the L4 level (high-level autonomous driving technology) of 5G+V2X, and supplements with a beyond visual range physical twin remote control system to solve the problem of remote control of driverless vehicles under sudden road conditions. At the same time, it cooperates with the driver behavior and state perception detection system to monitor the driving decisions and physiological index changes of the driver in actual operation, and through simulation platform simulation, provides a theoretical basis for daily operation and pricing strategies, and builds a "human-vehicle-road-network" four-in-one short-distance driverless shuttle bus operation system; the structure is as Figure 1 shown;

[0024] Example 1: Autonomous driving sharing travel platform software

[0025] The present invention develops an autonomous driving shared travel platform software mainly based on carpooling on WeChat Mini Programs as the carrier, comprehensively calculates users' car-hailing requests, and plans the scheduling routes of autonomous driving minibuses; the platform software includes functions such as location selection, route planning, fare calculation, order adjustment, and user authentication, and has passed the test of the WeChat development platform, providing a good user experience; this platform is developed based on the JS, WXML, and CSS languages under the WeChat framework; on the starting page of the mini program, users select the starting point and destination by swiping the screen map; during this process, the platform accurately identifies various facilities based on the reverse geocoding and geocoding functions, and combines with the high-precision map to provide accurate and rapid guidance for the scheduling of autonomous driving minibuses; the platform provides a location search function based on the keyword search function, users can switch the regional list according to their needs, freely adjust the search range, and quickly select the location by supplementing with historical records and keywords to achieve an intelligent usage experience; after the location is selected, the platform obtains the coordinates of the shortest path point string in the driving mode based on the online map API and displays the route according to the map zoom factor; when used in conjunction with autonomous driving minibuses, the platform assists in scheduling based on data such as the head direction, speed, and acceleration in the vehicle package (BAG) data to achieve optimal route planning; the platform calculates the ride fare based on the starting price + mileage and supports subsequent adjustment of the calculation method according to the operating costs of autonomous driving minibuses; after the order is confirmed, the platform will display the estimated arrival time of the vehicle, and users can freely adjust or cancel the order during the waiting period; in order to provide better travel services, the platform also provides a user authentication function; after users log in, they can obtain historical order records and perform personalized settings; user request data is also extracted from the cloud background for analysis; the module architecture diagram is as shown in Figure 2 shown.

[0026] Example 2: High-precision map drawing, intelligent roadside device perception and visualization

[0027] High-precision map drawing and intelligent roadside equipment deployment are necessary conditions for building a community-based high-access autonomous travel system. In order to support intelligent connected vehicles to travel along established roads, the present invention collects and processes high-precision map data of roads, uses data collection vehicles equipped with sensors such as cameras, lidars, GPS, and IMUs to collect data on roads, and then synthesizes the lidar point cloud data, draws maps using the ArcGIS platform, and annotates road information such as lane lines and intersections. After the drawing is completed, a real vehicle deployment test is carried out, and the BAG data of the test vehicle during operation is used to verify whether the map is accurate, and the high-precision map is regularly maintained after the system is put into operation. Based on the above-mentioned high-precision map, the present invention constructs a twin base of a stereoscopic visualized digital road space, and uses roadside sensing equipment such as millimeter-wave radars, lidars, and smart cameras and vehicle-road collaborative facilities to build a smart traffic management platform. The platform uses sensing equipment to monitor road information in real time, synchronizes data with the platform through a 5G network, and visualizes the sensing data, and finally realizes holographic intersections and digital twins at the remote end. Its module architecture is shown in the figure below. Figure 3 shown.

[0028] Example 3: Autonomous driving minibus operating environment perception and trajectory prediction

[0029] In order to realize highly automated driving of minibuses, the present invention collects high-precision maps and uses them as the basis of the location information of the minibuses. At the same time, multimodal perception data is obtained through on-board perception devices such as lidar, millimeter-wave radar, and 360-degree surround camera to obtain real-time environmental information. On the basis of the prediction of the trajectory of the autonomous driving vehicle and surrounding targets, combined with real-time map information, collaborative perception of the autonomous driving minibus is realized. During the operation of the autonomous driving minibus, a BAG data packet is generated every 5 minutes, including 33 types of data such as inertial measurement unit data, autonomous driving status, millimeter-wave radar data, and information on moving objects beside the vehicle. By parsing the six-axis IMU data, three-dimensional linear acceleration and angle data with a frequency of 100Hz are obtained. By converting the data into displacement information and importing it into the six-axis single-person driving The simulator can simulate and reproduce the dynamic state of the driving process of the autonomous minibus; based on data analysis, the acceleration and speed change ranges are determined and candidate trajectories are generated through the collected vehicle status and moving object information beside the vehicle, and the trajectory data is classified through principal component analysis, K-means clustering algorithm and other methods; based on the large language model, a cross-domain transfer learning framework is designed to align the intra-domain features of the target domain intersection and the inter-domain features of the source domain intersection and the target domain intersection, and the trajectory is selected and output based on situation judgment and speed selection; combined with the driving simulation platform, the obtained trajectory is corrected by experts to realize the LLM-based migration and interpretation of trajectory prediction for the autonomous driving vehicle and the surrounding moving targets, and finally provide guarantee for the autonomous driving of the autonomous minibus; the module architecture diagram is shown in the figure Figure 4 shown.

[0030] Example 4: Beyond-line-of-sight physical twin remote control

[0031] To address emergencies that may occur and that autonomous minibuses are unable to handle on their own (such as temporary road closures, malfunction of a certain sensor, etc.), the present invention constructs a beyond-line-of-sight physical twin remote control module. By measuring the 5G signal strength on the main road, a visualized heat map of the 5G signal is drawn, and coordination with the operator is carried out for base station addition and signal reinforcement to provide a stable 5G communication environment, which serves as the digital foundation for the physical twin. During operation, in response to complex road traffic conditions, the on-vehicle control system of the autonomous minibus predicts the trajectories of surrounding moving targets based on multi-modal perception data and combines with a high-precision map, and determines through the system whether the current traffic situation can be resolved autonomously by the minibus. When the autonomous minibus is unable to resolve the encountered emergency on its own, the on-vehicle system uses 5G high-speed communication to transmit remote videos and images to the cloud control platform and displays them in the laboratory beyond-line-of-sight physical twin system. The operator in the laboratory uses the instant remote images, combines with the status of multi-modal on-vehicle perception data, and uses the beyond-line-of-sight physical twin system for real-time operation, thus realizing the temporary takeover and remote control of the autonomous minibus. Its module is as Figure 5 shown.

[0032] Example 5: Establish a perception and detection system for the status and behavior of passengers and drivers

[0033] Passengers and remote drivers are important participants in the autonomous driving system, and their perception of the vehicle operation status deserves great attention. Regarding the problem of driver status perception, multi-spectral information such as visible light and near-infrared is fused to establish a common perception technology for the limb postures, head orientations, and visual attention of drivers under different lighting, operation, and traffic environments. Combining millimeter-wave radar and wearable devices, the fusion perception of physiological characteristic information such as the electromyogram, heart rate, and respiration of drivers is realized. By integrating the advantages of devices such as multi-spectral vision sensors and millimeter-wave radar sensors, a multi-sensor multi-dimensional multi-modal driver status acquisition system is established and combined with the beyond-line-of-sight driving system, aiming to achieve real-time acquisition of personnel status under natural driving conditions and high-precision data construction. For passengers of autonomous driving services, the present invention intends to use multi-dimensional physiological indicators to comprehensively focus on the feelings of passengers during the autonomous driving process, and combine the first-person micro-experience of passengers with the actual operation scenario to study the synchrony between passengers, the synchrony between passengers and the vehicle, and the synchrony between the behaviors of passengers and remote drivers. At the same time, compared with the traditional riding experience, the differences brought by autonomous driving to the micro-feelings of passengers and drivers are also specifically studied.

[0034] In Figure 1In order to enhance the user's intelligent travel experience, the present invention combines an information service cloud control platform to build a "Mobility as a Service" community-based high-access autonomous travel system with functions such as one-stop travel, reservation-based travel, autonomous service, instant demand response, and dynamic supply-demand matching. The system fully considers the travel needs of passengers, takes the integrated cloud control platform of safety management and application services as the core, uses 5G+L4-level electric autonomous driving minibuses as the carrier, takes high-precision maps as the digital base, and is supplemented by intelligent roadside devices as the support to realize the operation of a four-in-one collaborative autonomous driving short-distance shuttle bus of "people-vehicle-road-network".

[0035] In Figure 2 this module, a software for an autonomous driving shared travel platform mainly based on carpooling is developed on a WeChat mini-program. It comprehensively calculates the user's car-hailing request and plans the scheduling route of the autonomous driving minibus, aiming to provide a low-carbon and convenient travel service.

[0036] In Figure 3 this invention, by collecting and processing laser point cloud data, drawing high-precision maps and conducting log inspections, serves as the twin base for constructing a stereoscopic digital road space. A smart transportation management platform is constructed using intelligent roadside perception devices and vehicle-road coordination facilities, and finally, holographic intersections and digital twins are realized remotely.

[0037] In Figure 4 this invention, high-precision maps are used as the basis for location information, and multi-modal perception data is obtained through in-vehicle devices to obtain real-time environmental information. Based on the prediction of the trajectories of autonomous driving vehicles and surrounding targets, combined with real-time map information, collaborative perception of autonomous driving minibuses is realized.

[0038] In Figure 5 this invention, when the unmanned minibus cannot solve emergencies autonomously, remote video transmission of the cloud control platform is used and displayed in the ultra-long-range physical twin system. Combining the status of multi-modal in-vehicle perception data, remote real-time operation is carried out using the ultra-long-range physical twin system to realize the temporary takeover and remote control of the minibus.

[0039] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An autonomous travel system based on artificial intelligence and autonomous driving technology, characterized by: The travel system is based on WeChat applet and autonomous driving shared travel platform software based on carpooling. It comprehensively calculates user ride requests and plans the dispatch routes of autonomous driving minibuses. The platform software includes location selection, route planning, cost calculation, order adjustment and user authentication functions, and has been tested on the WeChat development platform; the platform identifies various facilities based on reverse address resolution and address resolution functions, and combines maps to provide rapid guidance for the dispatch of autonomous minibuses; the platform provides location retrieval functions based on keyword search functions, and users can switch region lists according to their needs, freely adjust the search scope, and quickly select locations through historical records and keyword completion; after the location is selected, the platform obtains the shortest path point string coordinates in driving mode based on the online map API, and displays the path according to the map zoom factor; When used in conjunction with an autonomous minibus, the platform assists in scheduling based on the vehicle's head direction, speed and acceleration data in the vehicle package data to achieve optimal path planning;.

2. The autonomous travel system based on artificial intelligence and autonomous driving technology according to claim 1, characterized in that: The platform calculates the fare based on the starting price + mileage, and supports subsequent adjustments to the calculation method based on the operating costs of the autonomous minibus; after the order is confirmed, the platform will display the estimated arrival time of the vehicle, and users can freely adjust or cancel the order during the waiting period; in order to provide better travel services, the platform also provides user authentication functions; after logging in, users can obtain historical order records and make personalized settings; user request data is also extracted from the cloud background.

3. The autonomous travel system based on artificial intelligence and autonomous driving technology according to claim 1, characterized in that: The system also includes a beyond-line-of-sight physical twin remote control module; by measuring the 5G signal strength on trunk roads, drawing a 5G signal visualization heat map, and coordinating with operators to add base stations and reinforce signals, a stable 5G communication environment is provided, which serves as the digital foundation of the physical twin.

4. The autonomous travel system based on artificial intelligence and autonomous driving technology according to claim 3, characterized in that: During operation, in response to complex road traffic conditions, the on-board control system of the autonomous driving minibus predicts the trajectories of surrounding moving targets based on multimodal perception data and maps, and determines through the system whether the current traffic conditions can be resolved autonomously by the minibus; when the autonomous driving minibus is unable to autonomously resolve the emergency situation, the on-board system will use 5G high-speed communication and the cloud control platform to remotely transmit video and images, and display them in the laboratory's beyond-line-of-sight physical twin system; operators in the laboratory use real-time remote images combined with the status of multimodal on-board perception data to perform real-time operations using the beyond-line-of-sight physical twin system.

5. The autonomous travel system based on artificial intelligence and autonomous driving technology according to claim 1, characterized in that: By using visible light and near-infrared, the driver establishes a common perception technology that adapts to the driver's body posture, head orientation and visual attention in different lighting, working and traffic environments; combined with millimeter-wave radar and wearable devices, the driver's electromyography, heart rate, and respiratory physiological characteristics information can be integrated and perceived.

Images