Method for navigating a mobile robot in an urban environment

Abstract

FIELD: robotics.SUBSTANCE: control systems for mobile robots performing navigation tasks in an urban environment. A method for navigating a mobile robot in an urban environment consists of using an information system containing a sensor module equipped with at least a digital camera, a lidar, a radar, an inertial module, and a time synchronization module. In this case, the people are determined in the robot’s movement zone, their position and speed, as well as the orientation of the body and the direction of their gaze, and form a directional field of social comfort around each person, the parameters of which depend on the specified orientation, direction of gaze, and social status. A map of movement restrictions and preferences is formed as a combination of the specified fields and other obstacles, and the trajectory and speed of the robot's movement are selected based on this map. If the reliability of human recognition decreases or the robot becomes partially invisible, the field of view in the sector of the person's probable movement is expanded, the speed is limited, or the robot is stopped, the history of movements and events in the specified location is recorded, and it is used to update the social tension map. In this case, the generated information is transmitted to the robot via a communication channel and / or processed on board the robot.EFFECT: reduction in the number of social conflicts while maintaining the robot's throughput.6 cl, 1 dwg

Description

[0001] Technical field

[0002] The invention relates to the field of robotics, namely to control systems for mobile robots performing navigation tasks in an urban environment.

[0003] Technology Level

[0004] A prior art discloses a method for navigating a mobile service robot (patent RU2736559C1), comprising a movable platform, sensors with a controller, wheel motors with a controller, a human-machine interface unit, a communication unit, and a service device, as well as a navigation system consisting of a computer installed on a mobile MSR platform, designed for navigation using maps and inventory and various modules, including a navigation module and a rack identification module. The method includes: obtaining a static map of the CO, constructing a global map by scanning with sensors, and determining a coordinate system for the CO on the global map. The input data includes: a static map, a diagram for dividing the static map into zones, a sequence of zones, a zone scanning direction, and a rack inspection direction.The navigation method includes the following stages: specifying the first scanning area; forming a bypass scheme; generating waypoints; receiving input data from the rack identification software module; if a contour is detected, it is analyzed for study; if not detected, the condition for completing the scanning of the current zone (CZ) and the presence of the following zones for scanning are checked; if the contour has not been surveyed, proceed to task formation: the bypass stage; if the contour is known, the condition for completing the CZ scanning and the presence of the following scanning areas are checked; if the CZ scanning is completed and no other zones for scanning are available, the MSR returns to its initial position.

[0005] A drawback of existing solutions is the lack of consideration of people's social status and dynamic personalization of their "comfort zone." Anisotropic fields dependent on posture and gaze direction are not generated; a long-term "social tension map" of the location is not used for predictive behavior modification; two-way synchronization of intentions via eHMI and deterministic fail-safe rules are lacking in the event of a drop in recognition reliability.

[0006] Disclosure of the essence of the invention

[0007] The objective of the invention is to ensure safe and efficient movement of a robot in an urban environment.

[0008] The technical result consists in eliminating conflict interactions with people and dangerous approaches while maintaining the robot's throughput.

[0009] The method for navigating a mobile robot in an urban environment is characterized by the use of an information system comprising a sensor module equipped with a digital camera, lidar, radar, inertial unit, and time synchronization module. The "information system" refers to the robot's onboard system or an external infrastructure system with wireless data exchange.

[0010] The information system determines the spatial position and speed of people within the robot's range. It also identifies static and dynamic objects that interfere with the robot's movement and the time until the robot approaches a person or artificial obstacle. It then generates a directional (anisotropic) social comfort field around each person and takes into account the location's social tension map.

[0011] The robot's trajectory is generated, its speed and stopping point are determined, and data about the robot's movements in a specific location is recorded in the information system's memory. Based on the recorded events, a social tension map of the location is generated and updated; the contribution of each event decreases over time with a specified constant.

[0012] A safety zone is created for the robot's movement when a person is not fully visible, or the safety zone is expanded in the direction of possible human movement. Using the information system, the generated information is transmitted to the robot via a communication channel.

[0013] Sensor module - provides dynamic data about the environment and people.

[0014] People Condition Assessment Module - performs detection and tracking of people, assessment of posture, body orientation and gaze, classification of vulnerable road users (children, elderly people, people with strollers), as well as grouping of people and flow density assessment.

[0015] Social Comfort Field Generator - creates an anisotropic field around each person as a superposition of oriented functions (frontal, lateral, and rear components). The field's parameters depend on the person's state and can be adjusted based on biometric indicators.

[0016] The social stress map is a discrete grid with a 0.5-1.0 m step, where events such as sudden braking or forced stops accumulate. The contribution of each event decays exponentially over 24-72 hours.

[0017] Local planner is an optimizer that calculates the robot's trajectory while observing motion invariants, which are functions of local values.

[0018] The safety module imposes movement restrictions or initiates a stop in the event of a collision with humans, independent of the local planner. The eHMI external interface provides two-way synchronization of intentions with humans: light indicators, icons, and / or projections are used to clearly communicate the robot's status (going / waiting / yielding / stopped), with the ability to confirm via gestures or gaze. The interface is synchronized with the robot's movement dynamics and is updated with a delay of no more than 100 ms. The safety zone radius is set at 1.0-2.0 m; in cases of poor visibility, the zone is expanded by 0.5-2.0 m in the uncertainty sector.

[0019] The speed of approaching a person is limited to no more than 6 km / h, and in the safety zone - 2-3 km / h.

[0020] Recalculation of the trajectory in the presence of people is performed at a frequency of at least 5-10 Hz.

[0021] A method for navigating a mobile robot in an urban environment, characterized in that it uses an information system containing a sensor module equipped with at least a digital camera, a lidar, a radar, an inertial module and a time synchronization module, wherein: people in the robot's movement zone are determined, their position and speed, as well as the orientation of the body and the direction of view; based on the observed features, each person is assigned to a social state, for example, "attentive", "busy" or "vulnerable"; a directional field of social comfort is formed around each person, the parameters of which depend on the specified orientation, direction of view and social state;The social tension map of the location is taken into account, in which events characterizing conflict situations during movement (forced stops, abrupt braking, dangerous approaches) are accumulated, with the contribution of such events decreasing over time, and the parameters of the fields and weights used in movement planning are modified based on the values ​​of the specified map; a map of movement restrictions and preferences is formed as a combination of the specified fields and other obstacles, and the trajectory and speed of the robot are selected based on this map; if the reliability of human recognition decreases or partial invisibility occurs, the field in the sector of probable human movement is expanded, the speed is limited, or the robot is stopped; two-way synchronization of intentions with people is ensured via the external eHMI interface, coordinated with the current actions of the robot; the history of movements and events of the specified location is recorded and used to update the social tension map.the generated information is transmitted to the robot via a communication channel and / or processed on board the robot.

[0022] The Social Tension Map stores events with decreasing contribution over time, with the nature of the decrease being determined by the location's "memory" time parameter, and takes into account the daily and weekly recurring features of the location.

[0023] When a person is partially invisible, the field in the corresponding sector is expanded by 0.5-2 m, and the basic radius of the field is set to 1-2 m.

[0024] The speed of approaching a person is limited to no more than 6 km / h, within the field the speed of movement is limited to 2-3 km / h, and the recalculation of the trajectory in the presence of people is performed at a frequency of no less than 5-10 Hz.

[0025] When the specified tension level in the social tension map is exceeded, the speed is reduced in advance and the trajectory is changed, bypassing such zones.

[0026] The generated map of restrictions and preferences additionally takes into account static elements of the urban environment and temporary traffic restrictions.

[0027] Brief description of the drawing

[0028] Fig. 1 shows a block diagram of the interaction of the modules of the information system

[0029] 100 - sensor module; 110 - people's condition assessment module; 120 - directional (anisotropic) field generator; 130 - social stress map ∑; 140 - trajectory and speed planner; 150 - safety module (fail-safe); 160 - eHMI external interface; 170 - robot; 180 - human.

[0030] Implementation of the invention

[0031] As the robot moves along the sidewalk, the information system uses a sensor module to detect the presence of people, objects, and their density. The information system then sets the robot's route to avoid collisions with people or obstacles. In real time, the information system detects changes in the surrounding environment and transmits information to the robot via a communication channel about changing its trajectory or stopping completely at its current location. All route data and related events are recorded in the information system's memory.

[0032] The eHMI interface provides communication with people: light indicators, floor projections, pictograms, audio signals, the interface is synchronized with the dynamics of movement and is updated with a delay of no more than 100 ms, the radius of the safety zone is set at 2 meters, in case of partial visibility, the zone is expanded by 2 meters in the uncertainty sector, the speed of approaching a person is limited to no more than 6 km / h, and in the safety zone - 2 km / h, the trajectory recalculation in the presence of people is performed at a frequency of at least 5 Hz.

Claims

1. A method for navigating a mobile robot in an urban environment, characterized by the fact that they use an information system containing a sensor module equipped with a digital camera, lidar, radar, inertial module and a time synchronization module, distinguished by the fact that they determine the position in space and the speed of movement of people in the area of ​​​​the robot's movement, determine static and dynamic objects that are an obstacle to the movement of the robot, determine the time until the robot approaches a person or an artificial obstacle, form the trajectory of the robot's movement and determine the speed of the robot and the place of its stop, record data on the movement of the robot in a certain location in the memory of the information system, form a safety zone in the absence of full visibility of a person or expand the safety zone in the direction of the possible movement of a person, with the help of the information system, the generated information is transmitted via a communication channel to the robot,The eHMI interface enables communication with people; the interface is synchronized with the dynamics of movement and is updated with a delay of no more than 100 ms; the radius of the safety zone is set at 1-2 m; in case of partial visibility, the zone is expanded by 0.5-2 m in the uncertainty sector; the speed of approaching a person is limited to no more than 6 km / h, and in the safety zone - 2-3 km / h; the recalculation of the trajectory in the presence of people is performed at a frequency of no less than 5-10 Hz.

2. A method for navigating a mobile robot in an urban environment, characterized in that they use an information system containing a sensor module equipped with at least a digital camera, a lidar, a radar, an inertial module, and a time synchronization module, wherein: they determine people in the robot's movement zone, their position and speed, as well as the orientation of the body and the direction of their gaze; they form a directional field of social comfort around each person, the parameters of which depend on the specified orientation, direction of gaze, and social state; they form a map of movement restrictions and preferences as a combination of the specified fields and other obstacles, and they select the trajectory and speed of movement of the robot based on this map; if the reliability of people recognition decreases or partial invisibility occurs, they expand the field in the sector of probable movement of the person, limit the speed, or initiate a stop of the robot;They record the history of movements and events in a given location and use it to update the social tension map; the generated information is transmitted to the robot via a communication channel and / or processed on board the robot.

3. The method according to paragraph 2, characterized in that, when a person is partially invisible, the field in the corresponding sector is expanded by 0.5-2 m, and the basic radius of the field is set at 1-2 m.

4. The method according to paragraph 2, characterized in that the speed of approaching a person is limited to no more than 6 km / h, within the field the speed of movement is limited to 2-3 km / h, and the recalculation of the trajectory in the presence of people is performed at a frequency of no less than 5-10 Hz.

5. The method according to paragraph 2, characterized in that when a specified level of tension in the social tension map is exceeded, the speed is reduced in advance and the trajectory is changed to bypass such zones.

6. The method according to paragraph 2, characterized in that the generated map of restrictions and preferences additionally takes into account static elements of the urban environment and temporary traffic restrictions.

Images