DE102023104142A1 - Orientation with external signals for mobile garden equipment - Google Patents

Abstract

The invention relates to a mobile gardening device, e.g. a robotic lawnmower or a lawn mower, and to a method for mapping the working environment of the gardening device and for locating the gardening device within the working environment. The gardening device (1) comprises a signal receiver (2) for detecting electromagnetic signals (S, S0, S1, S2, S3), in particular WiFi, Bluetooth, Zigbee, Thread or other signals, from one or more signal sources (e.g. from WLAN devices or smart home devices) from the environment. A map (M) is created and a position of the gardening device (1) is determined, wherein one or more signals (S) are detected, from which a signal fingerprint (SFP) is determined. The signal fingerprint (SFP) comprises a list (L) with one or more signal vectors (SV) comprising a signal identifier (Sn) and/or a signal intensity (Si) and/or a signal weight (Sw) of the detected signals (S). The signal finder print (SFP) is stored in association with a specific map position (PM).

Description

The invention is in the field of mobile gardening equipment, in particular robotic lawnmowers. The invention particularly relates to a mapping and localization method for mobile gardening equipment using signals already present in the working environment (e.g. from WLAN or Bluetooth devices).

The technology disclosed here is applicable to both autonomous mobile gardening equipment, e.g. robotic lawn mowers, and user-operated mobile gardening equipment, e.g. hand-pushed lawn mowers, lawn tractors. The technology is particularly advantageous for autonomous mobile gardening equipment such as robotic lawn mowers.

The disclosure includes a mobile gardening tool, a mapping method, and a localization method.

Various mobile gardening devices, such as robotic lawnmowers, are known from the state of the art and can move at least partially autonomously on a work surface, e.g. a lawn to be mowed.

Various technologies and methods are known from the state of the art for navigating gardening equipment, only some of which use a localization and mapping method. A simple and widely used method for controlling robotic lawnmowers is to drive a random path (so-called random walk) within a lawn area that is delimited by a lawn boundary wire. The lawn boundary wire is usually buried at a shallow depth in the ground. The robotic lawnmower uses sensors to detect an electromagnetic field that emanates from the lawn boundary wire to detect when the robot approaches the lawn boundary. When the lawn boundary is reached, the robotic lawnmower turns and drives a new random path within the lawn area until it encounters a lawn boundary again or encounters an obstacle. Various methods are also known for detecting and avoiding obstacles. With this simple navigation method, the robotic lawnmower does not know its position in the environment and drives on random paths until it accidentally encounters an obstacle or a lawn boundary, the positions of which in the environment it also does not know.

The state of the art also includes methods in which the robot can determine its position in the environment using various measurement methods. For example, robotic lawnmowers and navigation methods with satellite navigation (e.g. GPS, Galileo), real-time kinematics (RTK) with satellite navigation and local reference stations, and localization methods using triangulation or multilateration via radio beacons are known. Such beacons can be, for example, mobile radio transmitters that the user mounts at fixed locations in the work environment so that the robotic lawnmower can orient itself using their radio signals.

The known solutions have the disadvantage that the technology depends on additional hardware that the user has to place in the work environment solely for this purpose, or requires complex sensors for satellite navigation.

In some cases, the state of the art for mobile gardening equipment also uses navigation methods with maps to orientate them in the working environment. A map is a model representation of an environment (e.g. a lawn) that includes information about the location and nature of objects (e.g. lawn boundaries or obstacles). A map of a working environment has the advantage that the navigation of the mobile equipment can be adapted to the specific environmental conditions, e.g. to avoid known obstacles or to follow a particularly efficient path.

In order to plan a path related to the environment, the gardening tool needs localization (position determination). Localization can be global in relation to the world (e.g. via GPS) or local in relation to a specific lawn. In particular, position determination can be made in relation to a map of the local environment (e.g. the lawn). If a gardening tool knows where it is on a map through localization, it can adapt its movement to the position of other objects on the map, e.g. to avoid obstacles or to follow a particularly efficient path.

The invention aims to provide an improved technique for navigating a gardening tool in a working environment. The invention solves the problem with the features of the independent claims.

A computer-implemented method for mapping a work environment and localizing a mobile gardening device in the work environment is disclosed, in which one or more electromagnetic signals (e.g. radio beacon signals, WiFi signals or Bluetooth signals) are detected with a signal receiver of the gardening device and used for mapping and position determination. In addition, a corresponding mobile gardening device and a system with at least one gardening device and a server are disclosed. The method and device features disclosed below are to be understood in such a way that they can refer both as method features to the method and analogously as functional features to the devices. For simplicity, the invention is primarily described using the method.

The detected signals can come from various signal sources. The invention is based in particular on the advantageous idea of using signals that are already present in the work environment without necessarily having to arrange additional signal sources, e.g. beacons, in the work environment. Especially in urban environments, there are now many radio signal transmitters with mostly fixed positions, e.g. WLAN routers, Bluetooth devices, smart home devices or other radio signal transmitters that can be used to orient a gardening tool. This has the particular advantage that the mapping and positioning of a mobile tool can be improved using resources already available in the work environment.

A particular advantage of the invention is that the method can be implemented with little or even no additional hardware in existing gardening equipment, e.g. robotic lawnmowers. Many gardening devices already have signal receivers that can process signals from the most common signal standards (e.g. WiFi, Bluetooth, 433 MHz band, Zigbee, Thread). Depending on the design, the method can be implemented on existing gardening equipment through technical software updates. The functions can also be built or retrofitted with simple functional components.

It is particularly advantageous to use a signal receiver that can process signals from different types of radio signals (e.g. frequencies and modulations). This allows the technology to be used flexibly in as many different work environments as possible. In a particularly advantageous version, signals from external sources, i.e. sources independent of the gardening system (e.g. WLAN routers, Bluetooth devices, smart home devices) and additional signal sources created specifically for this purpose (e.g. beacons, radio transmitters of a base station) can be combined.

The signal receiver receives one or more signals and determines a signal fingerprint. The signal fingerprint represents an image of the signals recorded at a specific point in time or time interval at a specific location in the work environment and can include additional information. Assuming that at least some of the signal sources are stationary and the signals can be received over a longer period of time, information about the position of the signal receiver, i.e. the gardening tool, can be derived from the signal fingerprint.

The signal fingerprint includes at least information about the respective signal intensity (e.g. an RSSI value) of the received signal as well as a recognizable signal identifier (e.g. a WiFi SSID or a MAC address of the signal source). Advantageously, the signal fingerprint also includes a signal weight that is assigned to the received signal. The signal weight can, for example, contain information about the reliability of a specific signal in order to give more consideration to reliable signals in mapping and/or positioning than to unreliable signals. A high signal weight can, for example, be assigned to signals that have been recorded at the same location with the same intensity over a long period of time and therefore suggest a stationary signal source. A low signal weight could, for example, be assigned to signals that were detected for the first time, exhibit frequent interference, or were no longer detected at an expected location.

The signal fingerprint is stored in the map for mapping for a specific map position. This is preferably carried out during a mapping trip (also called a learning trip) for many, preferably all, positions in the map. Map positions can advantageously be re-mapped several times, in particular continuously throughout the entire operation of the gardening tool. Newly recorded signal fingerprints at a specific map position can be compared with previously recorded ones for the same or a similar map position. Depending on the design, new signal fingerprints can be accumulated or offset against each other (e.g. by adjusting the signal weights).

The signal fingerprints preferably comprise a list of signals. A signal vector can be stored for each signal. A signal vector refers to a coherent data structure with data that is assigned to a signal (e.g. a signal identifier, a signal intensity, a signal weight). The signal vector can be formed in any data format and is not restricted to a specific vector representation of a programming language or a mathematical representation. The list of signal vectors also represents only a collection of signal vectors that is not restricted to a specific order or data structure.

The signal fingerprint is used to create a recognizable data structure that can be linked to a map position and stored in the map. The map can be any data structure for representing an environment and/or topography. Preferably, a multidimensional discrete raster map is used. Other map formats are also possible. One or more signal fingerprints can be stored for each map position, which can later be used to determine the position. The signal fingerprints and/or map positions can also include information about objects in the environment (e.g. obstacles, lawn borders, base station).

Preferably, the method is used for simultaneous mapping of the working environment and position determination of the gardening tool. The method can therefore also be referred to as a method for simultaneous positioning and mapping (SLAM). During mapping, signal fingerprints are stored for specific map positions, which can be assigned to a position of the gardening tool to be determined at the time the signal fingerprint is recorded.

The method includes determining the position of the gardening tool. Depending on the embodiment, the position can be determined in several ways. When determining the position of the gardening tool, signal fingerprints previously stored in the map are advantageously also processed and combined with other positioning data (e.g. odometry data and/or boundary wire sensor data). To determine the position, a currently recorded signal fingerprint can in particular be compared with signal fingerprints previously stored in the map in order to determine similar or matching signal fingerprints and to determine at which map position they were previously stored.

To create and/or update the map, the gardening device preferably carries out one or more mapping trips, during which additional data is processed to determine the position of the gardening device. In particular, sensors for odometry (e.g. acceleration sensors, cycle path sensors) and/or sensors for detecting lawn boundary wires and/or other environmental sensors can be used for this purpose. In particular, user input data (e.g. for detecting the course of the lawn boundary) can be processed.

The more often the procedure is carried out for the different map positions (even multiple times for the same map positions), the more accurately the position of the gardening tool can be determined at any point in the work area using a combination of sensor data and signal fingerprints. This makes it possible to create a map that enables oriented path planning (instead of or in addition to the random walk). In addition, stationary obstacles that have been detected at least once at a specific map position and saved in the map can be avoided on future journeys using the map.

Further advantageous features and embodiments of the invention are described below with reference to the drawings.

• 1: eine schematische Draufsicht auf einen Mähroboter (1) in einer Arbeitsumgebung mit einer Basisstation (4), die durch einen Begrenzungsdraht (3) begrenzt ist und in deren Umgebung mehrere Signalquellen (Q) angeordnet sind;
• 2: die Bestimmung eines Signalfingerabdrucks (SFP) für drei empfangene Signale (S1, S2, S3) an einer bestimmten Kartenposition (PM).

The invention is illustrated by way of example and schematically in the drawings. These show:

• 1 : a schematic plan view of a robotic lawnmower (1) in a working environment with a base station (4) which is delimited by a boundary wire (3) and in whose surroundings several signal sources (Q) are arranged;
• 2 : the determination of a signal fingerprint (SFP) for three received signals (S1, S2, S3) at a specific card position (PM).

1 shows a simplified representation of a working environment for which a map (M) is created. The map can be simplified as a grid of discrete positions with a certain row and column position sition. Alternatively or additionally, three-dimensional maps or other environment representation models can be used as maps.

The invention is described using a robotic lawnmower as a mobile gardening device (1) as an example. The mobile gardening device (1) comprises a signal receiver (2, not shown). The signal receiver can comprise an antenna arrangement with one or more antennas as well as a processor and memory. The signal receiver is designed to receive one or more electromagnetic signals. Depending on the type of signals, the signal receiver can use the same or different antennas.

In the working environment there are usually one or more base stations (4) from which one or more boundary wires (3) extend. Alternatively or in addition to the boundary wires, the lawn boundary can also be determined by other methods (e.g. using optical sensors or user input).

As part of the method, the position of the gardening tool (1) is determined and a signal fingerprint is stored for this map position (PM). To determine the signal fingerprint, one or more signals (S0, S1, S2, S3) are recorded that originate from one or more signal sources (Q) in the environment. The signals can in particular originate from remote signal sources outside the work area (e.g. from a WLAN router in the house next to the lawn) and/or from signal sources in or near the work area, e.g. from a transmitter of the base station (4).

In a particularly advantageous embodiment of the method, a mapping run is first carried out along the lawn boundary (3). During this drive oriented along the lawn boundary, the position of the gardening tool can be determined in particular using the data from an odometry (e.g. distance measurement based on the wheel revolutions) and/or the data from a sensor for detecting the boundary wire. The signal fingerprints can be saved for these map positions on the lawn boundary and used for further mapping and work runs.

In a particularly advantageous embodiment, an additional mobile terminal, e.g. a smartphone of the user, can be arranged on the mobile working device during the mapping trip. Preferably, the mobile gardening device (1) or a connected server has a data communication interface to which the mobile terminal can be connected in order to obtain positioning data of the mobile terminal (e.g. the satellite navigation or inertial sensors of the smartphone) and to process them for the method disclosed here.

After a first mapping run along the lawn boundary, preferably at least one further mapping run is carried out within the work area along a random path (random walk). Preferably, the position determination for mapping in this phase is carried out on the basis of odometry data and possibly further position determination data. In a particularly advantageous embodiment, the position of the gardening device (1) is estimated using a state estimator, e.g. a Kalman filter.

To estimate the position of the gardening tool (1), a prediction of the current position is preferably carried out based on the measured self-motion in combination with the processing of the signal fingerprints for position determination, preferably starting from the last known position. In particular, odometry sensors and other navigation sensors can be used for this purpose. The accuracy of the sensor data is advantageously taken into account when estimating the position.

As soon as a map with sufficient signal fingerprints for the various map positions is available, the map can be used more for orientation of the gardening equipment and path planning. The map can now be used to plan and follow oriented paths instead of a random path.

For orientation within the map, signals (S) from the environment are continuously recorded and processed. 2 shows an exemplary determination of a signal fingerprint (SFP) from three received signals (S1, S2, S3). The first signal has a unique signal identifier (Sn) S1, a signal intensity (Si) of 10 and a signal weight of 5. The second signal has the signal identifier S2, a signal intensity of 87 and a signal weight of 7. The third signal has the signal identifier S3, a signal intensity of 36 and a signal weight of 3.

The values for signal identification, signal intensity and signal weight form a signal vector (SV) that represents the signal S1, S2 or S3 received at this position at this time. A list can be formed from the signal vectors that represents the signal fingerprint at this position.

Under otherwise identical conditions, the gardening tool would record the same signal fingerprint at the same position at a later point in time. The signal intensity preferably describes a standardized measure of the intensity of the signal, which depends on the distance of the signal receiver from the source of the signal. If the signal sources (Q) are stationary, transmit constantly and are not disturbed (e.g. by changing objects between the signal source and signal receiver), the position of the gardening tool can be determined from the changing signal intensities as it moves, since the distance to the respective signal sources changes accordingly. The method works better the more undisturbed and stationary signal sources are available to receive from different directions.

In a particularly advantageous embodiment, the received signals are assigned different signal weights, which are taken into account when processing the signals for position determination. A high signal weight is assigned to signals with high reliability and a low signal weight to signals with low reliability. The signal weight is advantageously continuously updated during operation of the method.

If the same signal is recorded repeatedly at the same map positions with the same signal intensity, this indicates a high level of signal reliability. This means that the signal is stationary and rarely disturbed. Conversely, signals that are recorded multiple times at the same position with different intensities have a lower level of reliability.

The signal fingerprints (SFP) are continuously compared with the stored signal fingerprints to determine the map position (PM) of the gardening tool (1). If the recorded signal fingerprint (SFP) is similar to a signal fingerprint previously stored for a specific map position, it can be concluded that the gardening tool is located at or near the map position for which this signal fingerprint was previously stored.

The map position (PM) determined using the signal fingerprints (SFP) can be offset against other positioning data, e.g. odometry, to improve positioning accuracy. Preferably, the data are fused in a state estimator.

In a particularly advantageous embodiment, the mobile gardening device can communicate with a remote server. The server can be accessible via the Internet (cloud server), integrated in the base station or in another gardening device. Parts of the method can be carried out on a data processing unit of the gardening device and/or the server.

Reference symbols

1 Garden robots Gardening robot 2 Signal receiver Signal receiver 3 Lawn edging Lawn boundary 4 Base station Base station 5 Data processing facility Data processing device 6 Control unit Control unit L list List M Map Map PM Map position Q Signal source Signal source S signal signal SFP Signal fingerprint Signal fingerprint Sn Signal identifier Signal identifier Si Signal intensity Signal intensity sv Signal vector Signal vector Sw Signal weight Signal weight

Claims (15)

Method for mapping a working environment and localizing a mobile gardening device (1) within the working environment, wherein the gardening device (1) comprises a signal receiver (2) which is designed to process electromagnetic signals (S) from one or more signal sources (Q) from the environment, comprising the following steps: - creating or obtaining a map (M); - determining the position of the gardening device (1); - detecting one or more signals (S, S0, S1, S2, S3); - determining a signal fingerprint (SFP), wherein the signal fingerprint (SFP) comprises a list (L) with one or more signal vectors (SV) of the detected signals (S, S0, S1, S2, S3), wherein the signal vector (SV) comprises a signal identifier (Sn) and/or a signal intensity (Si) and/or a signal weight (Sw) for a specific signal (S); - Storing the determined signal fingerprint (SFP) in connection with a specific card position (PM); Procedure according to Claim 1 , wherein the signal receiver (2) is designed to process signals according to at least one of the following standards: Wifi, Bluetooth, Zigbee, Thread, 433 MHz band. Procedure according to Claim 1 or 2 , wherein the list (L) is created and/or updated dynamically during the execution of the navigation method, in particular depending on the signals (S, S0, S1, S2, S3) present in the environment. Method according to one of the preceding claims, additionally comprising: - comparing the determined signal fingerprint (SFP) with one or more previously stored signal fingerprints (SFP'); - adapting a signal weight (Sw) of a specific signal source (Q) depending on the reliability of the associated signal (S). Method according to one of the preceding claims, wherein newly detected signals (S) are included in the list (L) and/or signals (S) are removed from the list (L) or ignored if a minimum reliability is not reached. Method according to one of the preceding claims, wherein the position of the gardening tool (1) is determined in relation to a lawn boundary (3) and/or a base station (4) and/or a reference signal transmitter. Method according to one of the preceding claims, wherein a mapping run is carried out along the lawn boundary, wherein in particular satellite navigation data, boundary wire sensor data and/or user input data are processed for determining the position during the mapping run along the lawn boundary. Method according to one of the preceding claims, wherein a mapping drive is carried out within the lawn boundaries along a random path. Method according to one of the preceding claims, wherein the position of the gardening tool (1) is determined on the basis of odometry data. Method according to one of the preceding claims, wherein the position of the gardening tool (1) is estimated by means of a state estimator, preferably by means of a Kalman filter. Method according to one of the preceding claims, wherein the position of the gardening tool (1), in particular during a mapping journey within the lawn boundary along a random path, starting from the last known position by a prediction using motion sensor data, preferably satellite navigation data, odometry data and/or their measurement error data, in combination with the detected signal fingerprint (SFP). Method according to one of the preceding claims, wherein obstacles are detected and stored in connection with a specific map position (PM). Mobile gardening device (1) comprising a signal receiver (2) which is designed to process electromagnetic signals (S) from one or more signal sources (Q) from the environment, a data processing unit (5) which is connected to the signal receiver (2), and a control unit (6) which is designed to control the gardening device (1), wherein the mobile gardening device is designed to carry out the method according to one of the preceding claims. Mobile gardening equipment according to Claim 13 , wherein the gardening tool (1) comprises one or more signal receivers (2), of which at least one signal receiver (2) is designed to process signals according to at least one of the following signal standards: 433 MHz band, Zigbee, Thread. System comprising at least one mobile gardening device (1) and a server, wherein the server is connected to the at least one mobile gardening device (1) via a data interface and the system is designed to carry out the method according to one of the preceding claims.

Images