Method for cluster formation

Abstract

The invention relates to a method and an automated entity for cluster formation. Automation entities are initially arbitrarily deployed on an operating area and called to be clustered to perform missions in an optimized and energy-efficient manner. The method comprises: (51) evaluating, by each automated entity, an initial individual level given its own communication capability and non-communication capability, and generating an individual planning strategy for reaching the convergence area under individual optimization constraints; communicating (52), by each automation entity, the evaluated initial level and capability data; and (53) establishing communication with the responding automation entities according to the initial levels and capability data received from the responding automation entities, and updating the initial individual levels and planning strategies to reach the convergence area while maintaining connection with the responding automation entities under cluster formation optimization constraints.

Description

Technical Field

[0001] The present invention relates to a method for distributed cluster formation, and relates to the field of connectivity of autonomous operating devices. Background Art

[0002] Many industries (e.g., agriculture, manufacturing, military, construction, logistics, etc.) believe that a higher level of autonomy is a means to improve operational efficiency and may also save resources such as energy, emissions, and personnel. This is accomplished through so-called "swarm intelligence", which means organizing a system of multiple automated entities to simulate the ability of a swarm to exhibit collective behavior in the face of external stimuli. For example, swarm farming is a new paradigm in agriculture, where small autonomous platforms in a swarm create new farming practices, as described in US2018020611A1 or US2019124910A1. Robots or aircraft can be used for handling dangerous missions, such as mine clearance (see CN112946766A) or disaster recovery (as disclosed in US2018217593A1). As disclosed in EP3545381A1, a swarm of micro air vehicles can perform autonomous surveillance. All of these solutions involve automated interaction between automated machines or vehicles and their orchestration. Thus, connectivity, or more precisely, mobile network coverage, is regarded as a key enabling factor for optimizing operational efficiency and allows for a high degree of flexibility.

[0003] Known connectivity solutions for off-road use cases vary widely in terms of application areas and requirements. Consider an off-road scenario with no network coverage: automated vehicles or robots, as members and communication nodes of a swarm, may operate in workplaces that are potentially huge in spatial size and have a large physical separation between each other. From the previous task, the nodes may end up in different locations that are potentially far apart. Due to the large physical separation, it is impossible to communicate directly with all nodes, and only some nodes in the entire swarm can communicate with each other. To obtain information related to the next task, the nodes need to create a fully connected swarm. The same need may also exist for antennas connected via non-ground networks (for connecting a spontaneously deployed wireless access network) or vehicle-to-everything communication (V2X, for sharing synchronization information), but for different reasons. A common solution to address this need is to have all nodes be "close enough" to a convergence point (e.g., in the center of the operation area). However, there are some drawbacks associated with this prior art: for example, many existing solutions can only be applied with the support of a mobile network, which can only provide limited coverage and may experience interruptions. Further, not every user or swarm operator wants to invest in deploying and maintaining a "private" local wireless access network.

[0004] Therefore, a cost-effective solution is needed that enables efficient operation of autonomous clustering even in areas without network coverage and in harsh environments. Here, direct or indirect communication between cluster members is beneficial and can also serve as a backup mode in the event of a failure of the local network infrastructure.

[0005] Therefore, the problem is how to efficiently create (fully) connected clusters while taking into account the capabilities of the wireless technologies used and considering efficiency optimization criteria such as the minimum time or minimum energy consumption required for the clustering formation of automated entities. Summary of the Invention

[0006] The object of the present invention is to provide such a method and corresponding automated entities.

[0007] This object is solved by a method for distributed clustering formation having the features of claim 1.

[0008] The dependent claims include advantageous further developments and improvements of the principles of the present invention as described below.

[0009] According to one aspect of the present invention, a method for distributed clustering formation of a plurality of automated entities initially deployed on an operating area is provided. At least one automated entity can be wirelessly connected to at least another automated entity among the plurality of automated entities. The method includes: each automated entity evaluating an initial individual rank given its own communication capabilities and non-communication capabilities and generating an individual planning strategy for reaching a convergence area under individual optimization constraints; each automated entity communicating the evaluated initial rank and capability data; and based on the initial rank and capability data received from the responding automated entities, establishing communication with the responding automated entities and updating the initial individual rank and planning strategy to reach the convergence area while maintaining connection with these responding automated entities under clustering formation optimization constraints.

[0010] Compared with the common method where all cluster members will converge at the same predefined location (e.g., usually at the center of the operating area) and all nodes are "close enough" to this location, a more advanced search strategy takes into account the capabilities of the cluster entities and allows optimization of the distributed search strategy for various objectives such as total energy or fuel consumption or individual energy or fuel consumption, the time to form or reorganize the entire cluster, etc.

[0011] In some embodiments, the communication capabilities can include communication range, transmission power, supported frequency bands, and wireless communication technologies such as radio, ZigBee, Bluetooth, WLAN, ultra-wideband, device-to-device communication, cellular or satellite communication. Additionally, the non-communication capabilities can include the type and characteristics of the automated entity, fuel or energy status, and on-board computing capabilities.

[0012] In a further embodiment, assessing an individual's rank means determining the priority and order of reaching the rendezvous area given the communication capabilities, distance to the rendezvous area or path characteristics, and the estimated fuel / energy or time consumption. Additionally, generating an individual planning strategy for reaching the rendezvous area under individual optimization constraints means determining one's own step size and / or trajectory considering individual time, space, or energy optimization constraints, and operating according to the behavior associated with the evaluated individual rank (i.e., as a scout or a follower of a scout). Further, updating the initial individual rank means adopting the order, priority, and behavior associated with the updated rank.

[0013] In some embodiments, adopting the behavior of a scout means that, while maintaining the (multiple) connections with the responsive automation entities and followers, each scout determines an updated step size and trajectory to reach the rendezvous area under optimization constraints. Adopting the behavior of a scout further means that if the scout estimates that it will lose its connection with its nearest (multiple) followers when moving according to the updated step size and trajectory, it shares the corresponding updated trajectory and step size with its followers. In other embodiments, adopting the behavior of a scout means that the scout periodically sends its own calculated step size and trajectory to the (multiple) followers. In yet another embodiment, adopting the behavior of a scout means periodically determining the step sizes and trajectories of itself and its followers and sending this data to the (multiple) followers.

[0014] In some embodiments, updating the individual rank further means that when two or more automation entities with similar ranks converge with each other, they compare their respective capabilities, and the entity with higher capabilities is assigned a higher rank.

[0015] In some embodiments, adopting the behavior of a follower means that the automation entity operates according to the step size and trajectory provided by the scout by correspondingly updating its own step size and trajectory. In some embodiments, adopting the behavior of a follower means that if the automation entity determines that it cannot follow the step size and trajectory provided by the scout due to its limited capabilities, the automation entity indicates its last position to the scout. Additionally, adopting the behavior of a follower further means forwarding any indication of the last position received from other followers to the scout.

[0016] In some embodiments, establishing communication means connecting with an automation entity, which is directly achieved using unicast or multicast through radio communication, device-to-device communication, 5G side-link, Wi-Fi direct, or ultra-wideband (UWB), through DSRC, LTE-V2X, NR-V2X, or geonetworked broadcast communication, or indirectly achieved in a multi-hop manner.

[0017] In some alternative embodiments, instead of having all cluster members arrive at the same unique rendezvous area, different clusters of automation entities are expected to arrive at different rendezvous areas while remaining connected under cluster formation optimization constraints.

[0018] According to another aspect of the present invention, there is provided an automated entity that is initially deployed arbitrarily on an operating area and is called to form a cluster. The entity includes a plurality of sensors, a wireless communication device, a memory, and at least one processor coupled to the memory. The at least one processor is configured to execute one or more instructions in the following items: evaluate the initial individual level based on its own communication capability and non-communication capability, and generate an individual planning strategy for reaching a convergence area under individual optimization constraints; communicate the evaluated initial level and capability data; establish communication with the responding automated entity based on the initial level and capability data received from the responding automated entity, and update the individual level and planning strategy to reach the convergence area, while maintaining connection with these responding automated entities under the cluster formation optimization constraints.

[0019] Here, the energy and fuel consumption of individual cluster members (automated entities) are also important parameters for the overall operational efficiency. In addition, communication costs and related trade-offs are also considered, especially when cellular or satellite communications are involved.

[0020] Further features of the present invention will become apparent from the following description and appended claims, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 An exemplary scenario is shown in which cluster entities are arbitrarily deployed over an operating area, a convergence area is defined, and automation entities are called to form a distributed cluster.

[0022] Figure 2 Shown Figure 1 The scene, in which Figure 1 The automation entities are shown connected in clusters, showing their respective wireless communication ranges,

[0023] Figure 3 Shows Figure 1 The self-assessment level of each automated entity participant in the scenario,

[0024] Figure 4 Presented participation Figure 1 The scene has been updated with the level of automation for dynamic movement of entities.

[0025] Figure 5 It is further shown that Figure 4 The scenario discussed in shows the dynamic progression of their approaching the convergence area and the corresponding communication range of each involved automation entity.

[0026] Figure 6 shows the moment when two scouts reach the rendezvous area and establish a fully connected cluster,

[0027] Figure 7 illustrates the possible ground topology of a fully connected cluster distributed in two rendezvous areas and supported by aerial entities,

[0028] Figure 8 shows the drone or aircraft part of a fully connected network of automated entities AE0–AE',

[0029] Figure 9 demonstrates a method for distributed cluster formation according to the present invention. DETAILED DESCRIPTION

[0030] To better understand the principles of the present invention, embodiments of the present invention will be explained in more detail below with reference to the accompanying drawings. In the drawings, like reference numerals are used for the same or equivalent elements and need not be described again for each drawing. It should be understood that the present invention is not limited to the disclosed embodiments, and the described features may be combined or modified without departing from the scope of the present invention defined in the appended claims.

[0031] As mentioned herein, "cluster automation entity" or "cluster entity" means an autonomous vehicle, robot, or unmanned aerial vehicle configured to simulate a cluster and capable of moving and operating under a common goal. Generally, once formed, the cluster movement is automatically coordinated among each other by a coordinator (operator or the automation entity itself), which is referred to as the "master entity" (or "cluster master entity"), and the term "slave entity" or "cluster slave entity" includes automation entities directly controlled by the master entity.

[0032] These automation entities should possess sensing, communication, and manipulation capabilities; for example, such automation entities can be equipped with several sensors (which implement object / obstacle detection and collision avoidance), such as cameras, radars, lidars, infrared, ultrasonic, etc. Additionally, devices for positioning are provided (e.g., GNSS sensors); functions such as vehicle movement (or flight of a UAV) are controlled (or programmed) by software (e.g., a drive-by-wire system). Furthermore, the same cluster entity can communicate using its on-vehicle communication unit that supports various radio technologies (e.g., radio, ZigBee, Bluetooth, WLAN, ultra-wideband, device-to-device, cellular, or satellite, but not limited to them).

[0033] The various embodiments described herein generally relate to communication techniques for the formation and self - coordination of clusters of automated entities that are called to achieve a common mission. However, as part of achieving dynamic cluster formation, the communication between automated entities must remain unobstructed regardless of how many levels in the cluster change, until the delegated mission is completed.

[0034] In the following, Figure 1 the exemplary scenarios shown are used to further explain how the method according to the invention operates.

[0035] Figure 1 An exemplary scenario is shown, in which an exemplary number of seven automated entities AE1, …, AE7 are arbitrarily deployed over an operating area OA, which can for example cover dozens or hundreds of hectares. The automated entities are activated and called to form a cluster entrusted with a mission, or they can follow a predefined cluster policy including mission type and task schedule and enter the cluster formation mode; inevitably, the initial large physical separation does not allow all automated entities to be fully connected, even though direct communication with one or several of them is possible. Another thing worth mentioning is that although the figures may show automated entities deployed on a ground plane, automated entities deployed in the air or as high - altitude mobile platforms can be enabled to supervise the cluster formation from a higher perspective.

[0036] To illustrate the inventive concept, a meeting area MA is shown within the operating area OA. It is worth mentioning that the meeting area can be any location within the operating area, it does not necessarily correspond to the center of the meeting area, and there can be not only one meeting area, but multiple meeting areas. A local map including the (multiple) meeting areas or the coordinates of the (multiple) corresponding meeting areas (e.g., GNSS coordinates) can be provided to each automated entity AE1, …, AE7. Initially, some of the automated entities closest to the meeting area MA, namely AE4 and AE5, are separated by a distance d sep that exceeds their respective wireless communication ranges. However, taking into account its own communication capabilities and range, each automated entity attempts to establish communication with at least one nearby automated entity. Further, each automated entity performs positioning and determines the distance or path to the indicated meeting area MA. Based on this data, each automated entity estimates the energy consumption required to reach the designated meeting area.

[0037] For a better understanding Figure 1 of the communication capabilities of Figure 2The corresponding wireless communication device ranges of each automated entity are shown by dashed circles. For example, AE3 and AE6 employ short-range communication ranges, while AE4 and AE5 are equipped with wireless communication devices with longer ranges. As depicted, each automated entity is a communication node, but not all nodes are within each other's coverage area: for example, entities AE2 and AE4 can communicate and form a first connected sub-cluster K1 (bounded by long double-dashed lines that overlap with the wireless communication range of AE4); AE5 and AE6 can form a second connected sub-cluster K2 (bounded by dashed lines) that overlaps with the communication range of AE5; while AE1, AE3, and AE7 are isolated, and the first cluster K1 and the second cluster K2 are separated. Those separated entities or sub-clusters can apply optimized search strategies to approach each other and (re)form an entire cluster.

[0038] Assume that the automated entities AE1, …, AE7 have completed the creation of the depicted topology for the previous task; in order to regroup to start another mission or task, the nodes need to (re)create a fully connected cluster. Compared with the common method where all cluster entities default to converge at the same predefined location (usually the center of the operation area OA), a more advanced strategy for forming a fully connected cluster takes into account the capabilities and knowledge of the cluster entities. Thus, under optimized constraints, based on predefined criteria (such as regarding mission type, entity type, communication capabilities, cluster composition in terms of individual fuel or energy or total fuel or energy), cluster formation objectives can be adopted, with optimized constraints such as: "minimum time to form a cluster", "minimum energy consumption to form a cluster", "minimum distance to form a cluster", etc.

[0039] As Figure 3 depicted, each automated entity can be programmed to determine its own initial rank within the cluster, which is represented by the following different parameters: distance to the convergence area, communication capabilities or range, energy status or fuel range, on-board computing capabilities. For example, when the rank is represented by the distance to the convergence area, the automated entities AE4 and AE5 have a first rank (high rank), AE2 and AE6 have a second rank, AE3 has a third rank, and AE1 and AE7 have a fourth rank (e.g., in a quantified form as RK1: d < 1 km, RK2: d < 2 km, RK3: d < 3 km, etc.). If an automated entity cannot reach the convergence area due to other limitations (such as fuel or energy shortage), the initial estimated rank is set to a lower value (e.g., AE3 from RK3 to RK4). In other words, low energy or low fuel means a downgrade in rank. This also applies to the remaining parameters used individually or in combination.

[0040] Once these data are estimated, each automated entity broadcasts at least its own identifier (ID), type (e.g., combine harvester or tractor), energy / fuel status, and initial rank, and monitors the wireless channel for possible response signals from other entities. Depending on changes in its status (e.g., an accident, an emergency, or an unexpected failure of a non-communication system that would prevent the automated entity from moving or operating as expected), additional data can be added to the broadcast signal: for example, a message "help needed" for a "search and rescue" operation.

[0041] In the case where an entity does not receive such a response within a predefined time limit due to its isolated status (e.g., the case of AE1, AE3, AE7 in the scenario of Figure 2 ), each entity determines its own step size and trajectory towards the convergence area MA and continues with an independent approach while continuously broadcasting its presence. For example, if an entity is already within the convergence area, it does not need to move, but it must still announce its presence and eventually connect in the cluster. If each automated entity receives a response from each other automated entity and meets the predefined cluster formation criteria, it means that a fully connected cluster has been formed. The cluster master entity (usually characterized by on-board computing capabilities) creates a mission plan and task schedule and distributes them to the cluster members.

[0042] In the following, a situation is described where an automated entity receives response signals from several other automated entities (e.g., each automated entity from two sub - clusters K1 and K2), but does not meet the cluster formation criteria. For example, automated entity AE2 can receive a response signal from AE4 and be able to communicate with it to exchange data; once communication is established within sub - cluster K1, the initial individual ranks are also exchanged and updated. If several automated entities are at the same distance within the convergence area, the entity with the highest energy or fuel status or capability assumes the role and corresponding behavior of a "scout". For example, since AE4 has a higher rank than AE1 and AE2 (at least according to the self - estimated initial rank), AE4 can implicitly assume the role and behavior of a scout within sub - cluster K1, i.e., while maintaining the connection of sub - cluster K1, lead the approach to the convergence area MA and search for other entities to be connected to the cluster. The same situation applies to sub - cluster K2, for which AE5 can assume the role of a scout. If there are still multiple entities within the same sub - cluster suitable for the role of a scout, the scout is determined randomly, e.g., using a contention resolution mechanism. Additionally, scout behavior means adopting one of the following options: a) If the scout estimates that it has lost connection with its (multiple) followers, it only emits its calculated step size and trajectory - the energy consumption level spent on communication is the lowest; b) The scout periodically emits its own calculated step size and trajectory to its (multiple) followers - a medium level of energy consumption; c) The scout periodically calculates the step size and trajectory data of itself and its followers and emits these data to its (multiple) followers - the calculation of all trajectories and their communication energy consumption is the highest.

[0043] Furthermore, any other entity that is connected to the corresponding scout and has a lower rank becomes a "follower". For example, AE4 has AE2 as a follower; AE6 is a follower of AE5. Adopting this "follower" behavior means that the corresponding automated entity can operate freely according to its initial planning strategy to reach the convergence area; moreover, the follower can operate according to the step size and trajectory provided by its corresponding scout and can update its own step size and trajectory accordingly. If a follower determines that it cannot follow the step size and trajectory provided by the scout due to its limited capabilities (e.g., energy / fuel status), the follower indicates its last position to the scout and continues to determine an alternative trajectory given the energy / fuel constraints. Additionally, each follower forwards any indication of the last position received from other followers to the scout.

[0044] Each scout AE4 and AE5 determines its "search step" according to specified optimization constraints or swarm formation goals. For example, if the optimization goal is to minimize the total energy or fuel consumption until the swarm formation is completed, the corresponding scout can only move within the range allowed by the communication range of its nearest follower and instruct other swarm entities to follow accordingly, but only to the extent necessary to maintain connectivity. To this end, the scout can estimate the distance to other responding entities, for example, based on the strength of the response signal or the time difference between two consecutive signals. Instead of estimating the distance based on radio signal ranging between swarm entities, the wireless communication quality can be considered. Thus, the "search load" is shared among the connected swarm entities.

[0045] Figure 4 The further dynamic movement of the updated-level automated entities when approaching the convergence area is presented. The initially isolated automated entity AE3 approaches the convergence area MA according to its own plan and is about to connect to the sub-cluster K1. Once it shares its data (as identifiers, levels, fuel or energy status, types, and other characteristics of this category), their respective initial levels will be re-evaluated. Updating the individual levels of two or more automated entities that converge with each other and have similar levels means that they compare their respective capabilities, and the entity with higher capabilities is assigned a higher level. For example, the updated level RK2 of the initially isolated automated entity AE1 is now higher than the initial level RK4 and higher than the initial level RK3 of another connected entity (i.e., AE2); the same applies to AE7, which is upgraded from RK4 to RK3 at this moment. Additionally, if a scout estimates that it may lose connection with its nearest (multiple) followers when moving according to the updated step and trajectory, the corresponding scout shares the corresponding updated trajectory and step with its followers. By doing so, even in the case where the corresponding scout leaves the communication range of one or more of its followers, the broadcast data (in addition to the updated step and planned trajectory, its own identifier, type, fuel / energy status, level, plus similar data received from the corresponding followers) will not be lost, and the corresponding followers can still use and forward or relay the scout's data, enabling all previously connected followers to follow the scout accordingly considering their own capabilities (e.g., energy / fuel status).

[0046] Figure 5 The further dynamic progress in the scenario discussed in Figure 4 is shown, demonstrating the corresponding communication ranges of each relevant automated entity.

[0047] Once the corresponding scouts AE4 and AE5 detect each other, they connect themselves and the previously separated sub-clusters K1 and K2. Once the two scouts come within communication range, they exchange information about their followers, such as their number, rank, fuel or energy status, and the characteristics of the swarm entities (vehicle type identifier). Further, they check if the swarm formation criteria are met. If so, they broadcast an indication that the swarm formation is complete and notify and trigger a predefined master entity to manage the swarm operations, such as mission configuration and creation and distribution of related task schedules. Otherwise, the two scouts consider the exchanged data, their individual capabilities, and predefined scout strategies to update their ranks to determine the new scouts. For example, the predefined scout strategy may instruct the first scout reaching the rendezvous area to wait at a predefined location for the other scout(s) to arrive and periodically emit discovery beacon signals.

[0048] Figure 6 Illustrates the moment when the scouts reach the rendezvous area and are fully connected. If it is determined that there are sufficient swarm nodes for the execution of the intended mission (meaning the swarm is complete and fully connected), the scouts report this to a predefined swarm master vehicle that oversees mission configuration and configuration and distribution of related tasks. Otherwise, the scouts can stay at the central location and wait for further scouts / nodes to arrive.

[0049] Figure 7 Illustrates a possible ground topology of a fully connected swarm distributed over two rendezvous areas and supported by aerial entities. Given the flexibility afforded by the fully connected swarm K, the swarm entities can be grouped for the various subtasks to be performed by the swarm without compromising the optimization of fuel / energy. For example, assume that the automated entities AE1 - AE3 prepare the soil; subsequently, the automated entities AE4 and AE5 then plant; and finally, the automated entities AE6 and AE7 complete the mission by fertilizing. This also applies to the grouping of AE 1’ –AE 7’ . Initially, the separation distance between the corresponding automated entities called to form the swarm exceeded their corresponding wireless communication range. Several rendezvous areas (MA1, MA2) are provided for the sub-clusters of connected entities: for example, AE1 - AE7 receive the first rendezvous area MA1 located at one end of the operation area; AE 1’ -AE 7’ receive a second rendezvous area MA2 different from the first rendezvous area. This rendezvous area definition can be part of the mission configuration created and distributed by the swarm master entity or part of a predefined mission strategy. Further, it is conceivable that aerial entities (e.g., drones or high-altitude mobile platforms) AE0 and AE 0’Support the positioning of cluster nodes and directly communicate and indicate to the cluster scouts and / or automated entities where to move for cluster formation.

[0050] Figure 8 Shows the drone AE0, as an automated entity AE Figure 7 from i 、AE i’ (i = 1,..., 7) as part of a fully connected network. AE 1’ 、AE 7’ 、AE1, and AE2 are within the communication range of AE0, which is represented by the distance d com .

[0051] To optimize their clustering and reach the designated rendezvous area in a timely manner with minimum fuel / energy consumption, at least one automated entity of each mentioned group can assume the role of a scout to search for other entities to connect to. For illustration, assume initially, AE0 is within the range of AE 1’ 、AE 7’ 、AE1, and AE2; AE 0’ is within the range of AE 3’ –AE 5’ . On the ground, AE i (i = 1,..., 7) are connected into a single cluster K, and the same is true for AE i’ (i' = 1,..., 7). Once the cluster formation is triggered, each relevant entity first evaluates its own initial rank given its own communication and non-communication capabilities - including AE0, AE 0’ . In this context, since the aerial entities AE0, AE 0’ can overlook the operation area OA globally, they can respectively assume the role of scouts and can approach the corresponding rendezvous area in the manner already described.

[0052] However, as long as the cluster formation criteria are not met (e.g., one of the planting entities is missing in the second rendezvous area), even if all automated entities are connected together, the cluster formation is considered incomplete. However, such missing entities can be easily detected within the fully connected cluster K without having to travel to a single rendezvous area and then return to travel to different rendezvous areas of different sub-connected clusters.

[0053] Figure 9 is a flowchart of the method according to the present disclosure.

[0054] According to a first aspect of the present invention, there is provided a method 100 for distributed formation of an automated entity cluster. A plurality of automated entities are initially deployed over an operating area; at least one automated entity is capable of wirelessly connecting to at least another one of the plurality of automated entities. Each automated entity called to form a cluster knows in advance one or more rendezvous areas. Once the cluster formation is triggered (started), the method includes:

[0055] Step 1: Each automated entity evaluates an initial individual rank given its own communication and non-communication capabilities, and generates an individual planning strategy for reaching the rendezvous area under individual optimization constraints.

[0056] Step 2: Each automated entity conveys the evaluated initial rank and capability data.

[0057] Step 3: Based on the initial rank and capability data received from the responding automated entities, establish communication with the responding automated entities, and update the initial individual rank and planning strategy to reach the rendezvous area while maintaining a connection with these responding automated entities under cluster formation optimization constraints.

[0058] Some embodiments provide that evaluating the initial individual rank means determining the priority and order for reaching the rendezvous area given communication capabilities, distance and path characteristics to the rendezvous area, and estimated fuel / energy or time consumption.

[0059] In some embodiments, communication capabilities include communication range, transmission power, supported frequency bands, and wireless communication technologies: radio, ZigBee, Bluetooth, WLAN, ultra-wideband, device-to-device, cellular, or satellite communication. Also included are variants such as UHF / VHF, Wi-Fi, LoRa, dedicated short range communication (DSRC), etc.

[0060] In some embodiments, non-communication capabilities include the type and characteristics of the automated entity, fuel or energy status, and on-board computing capabilities. Additionally, capability data may include data such as position relative to the operating area, relative orientation to other automated entities; sensor data (speed, acceleration, and / or relative speed relative to other automated entities), and operational data such as radio signal round-trip time estimation for ranging, radio connection to other automated entities, etc.

[0061] Generating an individual planning strategy for reaching a convergence area under individual optimization constraints means determining one's own step size and / or trajectory considering individual time, space, or energy optimization constraints, and operating according to the behavior associated with the evaluated initial individual rank (i.e., as a scout or a follower of a scout). Additionally, updating the initial individual rank means adopting the order, priority, and behavior associated with the updated rank. Further, updating the individual rank means that when two or more automated entities with similar ranks converge with each other, they compare their respective capabilities, and the entity with higher capabilities is assigned a higher rank.

[0062] Adopting the behavior of a scout means that, under the condition of maintaining connections with the responding automated entity(ies) and followers, each scout determines an updated step size and trajectory towards the convergence area under optimization constraints. Further, adopting the behavior of a scout means that if the scout estimates that it will lose the connection with its nearest follower(s) when moving according to the updated step size and trajectory, it shares the corresponding updated trajectory and step size with its followers.

[0063] Adopting the behavior of a follower means that the automated entity operates according to the step size and trajectory provided by the scout by correspondingly updating its own step size and trajectory. Additionally, adopting the behavior of a follower means that if the automated entity determines that it cannot follow the step size and trajectory provided by the scout due to its limited capabilities, the automated entity indicates its last position to the scout. Further, adopting the behavior of a follower means forwarding any indication of the last position received from other followers to the scout.

[0064] In some embodiments, establishing communication means connecting with an automated entity, directly using unicast or multicast through radio communication, device-to-device communication, 5G side link, Wi-Fi Direct or Ultra-Wideband (UWB), through DSRC, LTE-V2X, NR-V2X or geonetworked broadcast communication, or indirectly in a multi-hop manner.

[0065] Some embodiments may provide that different clusters of automated entities are expected to reach different convergence areas while maintaining connections under cluster formation optimization constraints.

[0066] Another aspect provided by the present invention is an automated entity that is initially arbitrarily deployed on an operation area and is called to (re)form a cluster to perform a mission. Such an entity may include multiple sensors, a wireless communication device, a memory, and at least one processor coupled to the memory. The at least one processor is configured to execute one or more of the following instructions: evaluate an initial individual rank based on its own communication and non-communication capabilities, and generate an individual planning strategy for reaching a convergence area under individual optimization constraints; communicate the evaluated initial rank and capability data; establish communication with a responding automated entity according to the initial rank and capability data received from the responding automated entity, and update the individual rank and planning strategy to reach the convergence area while maintaining a connection with these responding automated entities under cluster formation optimization constraints.

[0067] After being assigned the role of a scout, an automated entity calculates its own search step size and determines a trajectory or path towards the convergence area, depending on the condition of maintaining a connection with the responding automated entity and the follower(s) under the cluster optimization objective. The scout behavior may further include one of the following alternatives: a) if the scout estimates that it loses connection with the follower(s), it only emits its calculated step size and trajectory; b) the scout periodically emits its own calculated step size and trajectory to the follower(s); c) the scout periodically calculates the step size and trajectory data of itself and the follower(s) and emits this data to the follower(s). In other words, in some embodiments, under the cluster optimization objective, the scout may be able to periodically determine the individual follower step size and trajectory of each of its followers towards the convergence area according to the corresponding follower rank, and in addition, periodically communicate the correspondingly calculated step size and trajectory to each of its followers. In some embodiments, the scout may estimate that it loses connection with its (multiple) nearest followers when moving according to the determined step size and trajectory; in this case, the scout communicates its own search step size and trajectory to the followers. The choice of one alternative over the other depends on the cluster optimization objective, especially when it comes to the energy situation.

[0068] After being assigned the role of a follower, the automated entity may operate following the initially planned step size and trajectory, or may follow the step size and trajectory provided by the scout by correspondingly updating its own step size and trajectory. In some embodiments, if the follower determines that it is unable to follow the provided step size and trajectory due to its limited capabilities, the follower may indicate its last position to the scout. Additionally, the follower may forward any indication of the last position received from other followers to the scout, and the corresponding scout records any indication of the last position provided by the followers.

[0069] If any automated entity does not receive a response from any other automated entity among a plurality of automated entities, the automated entity operates autonomously according to an individual planning strategy.

[0070] The automated entities described herein include, for example, robots, automated guided vehicles, unmanned and autonomous systems, autonomous vehicles, industrial, agricultural or construction equipment, unmanned aerial vehicles, drones or high-altitude mobile platforms.

[0071] Additional potential use cases of the present invention are listed below: agricultural machinery and robots; mining; construction and road works; handling of goods, containers and supplies by autonomous automated guided vehicles (aAGVs).

[0072] Additionally, the present invention provides a cluster of automated entities, which is programmed according to the method of the present invention. The automated entities are initially deployed arbitrarily over an operating area and are called to (re)program the cluster in order to perform a mission. The cluster includes at least one automated entity as a scout and a plurality of automated entities as followers, and any one of the cluster entities can behave as provided by the method of the present invention. The cluster of automated entities can include the automated entities described herein, including combinations thereof, wherein aerial entities can support ground entities in a distributed formation. For example, drones, high-altitude mobile platforms or aircraft can assist in connecting the automated entities of the cluster, acting as communication repeaters or mobile base stations.

[0073] Furthermore, a computer-readable medium for distributed cluster formation of automated entities and storing computer-executable instructions is provided. As described above, the various functions or operations associated with the method disclosed herein can be implemented in one or more processors or computer systems operating in the automated entity. As is well known in the art, a general-purpose computer typically includes a central processing unit or other processing device, an internal communication bus, various types of memory or storage media for storing code and data (RAM, ROM, EEPROM, cache memory, disk drives, etc.), and one or more network interface cards or ports for communication purposes. Software functions involve programming, including instructions or executable code and associated stored data, for example, files for implementing various operations (including the operations or functions described herein according to the present disclosure). Software code can relate to client or server or network element functions and can be executed by a general-purpose computer. In operation, as described above, the code is stored in a non-transitory machine-readable storage medium within the computer platform. However, at other times, the software can be stored elsewhere and / or transmitted to a suitable general-purpose computer system for loading in order to execute. Software code for applications or other programming related to the operations and / or functions disclosed herein can also be stored in a server and transmitted over a network for storage in the memory of a client.

Claims

1. A method for clustering formation (100) of a plurality of automated entities initially deployed on an operating area, wherein, At least one of the plurality of automated entities is capable of wireless connection with at least another automated entity, and the method is characterized in that it includes: (S1) Each automated entity evaluates an initial individual rank given its own communication capabilities and non-communication capabilities, and generates an individual planning strategy for reaching a convergence area under individual optimization constraints. (S2) Each automated entity conveys the evaluated initial rank and capability data. (S3) Based on the initial rank and capability data received from the responding automated entities, establish communication with the responding automated entities, and update the initial individual rank and planning strategy to reach the convergence area while maintaining connection with these responding automated entities under cluster formation optimization constraints.

2. The method according to claim 1, wherein Communication capabilities include communication range, transmission power, supported frequency bands, and wireless communication technologies such as radio, ZigBee, Bluetooth, WLAN, ultra-wideband, device-to-device communication, cellular or satellite communication.

3. The method according to claim 1, wherein, Non-communication capabilities include the type and characteristics of the automated entity, fuel or energy status, and on-board computing capabilities.

4. The method according to claim 1, wherein (S1) Evaluating the initial individual rank means determining the priority and order for reaching the convergence area given communication capabilities, distance to the convergence area, path characteristics, and estimated fuel / energy or time consumption.

5. The method according to claim 1, wherein, (S1) Generating an individual planning strategy for reaching the convergence area under individual optimization constraints means determining its own step size and / or trajectory considering individual time, space, or energy optimization constraints, and operating according to the behavior associated with the evaluated initial individual rank, i.e., as a scout or a follower of a scout.

6. The method according to claim 1, wherein, (S3) Updating the initial individual rank means adopting the order, priority, and behavior associated with the updated rank.

7. The method according to claims 5 and 6, wherein, Adopting the behavior of a scout means that each scout determines an updated step size and trajectory towards the convergence area under optimization constraints while maintaining connection with the responding automated entities and followers (if any).

8. The method according to claim 7, wherein Adopting the behavior of a scout further means that if the scout estimates that it will lose connection with its nearest follower(s) when moving according to the updated step size and trajectory, it shares the corresponding updated trajectory and step size with its follower(s).

9. The method according to claim 7, wherein Adopting the behavior of a scout means that the scout periodically sends its own calculated step size and trajectory to the follower(s).

10. The method according to claim 7, wherein Adopting the behavior of a scout means that the scout periodically calculates the step size and trajectory data of itself and the followers, and sends this data to the follower(s).

11. The method according to claims 5 and 6, wherein, (S3) Updating the individual rank further means that when two or more automated entities with similar ranks converge with each other, they compare their respective capabilities, and the entity with higher capabilities is assigned a higher rank.

12. The method according to claims 5 and 6, wherein, Adopting the behavior of a follower means that the automated entity operates according to the step size and trajectory provided by the scout by correspondingly updating its own step size and trajectory.

13. The method according to claims 5 and 6, wherein, Adopting the behavior of a follower further means that if the automated entity determines that it cannot follow the step size and trajectory provided by the scout due to its limited capabilities, it indicates its last position to the scout.

14. The method according to claims 5 and 6, wherein, Adopting the behavior of a follower further means forwarding any indication of the last position received from other followers to the scout.

15. The method according to claims 5, 6, and 7, wherein, Adopting the behavior of a scout further means recording any indication of the last position provided by the followers.

16. The method according to claim 1, wherein (S3) Establishing communication means connecting to an automated entity, which is directly achieved using unicast or multicast through radio communication, cellular or satellite communication, device-to-device communication, 5G side link, Wi-Fi Direct or Ultra-Wideband (UWB), via DSRC, LTE-V2X, NR-V2X or geonetworked broadcast communication, or indirectly achieved in a multi-hop manner.

17. The method according to claim 1, wherein Different clusters of automated entities are expected to reach different convergence areas while maintaining connectivity under the constraints of cluster formation optimization.

18. An automated entity that is initially arbitrarily deployed over an operating area and is called to form a cluster, the entity comprising: A plurality of sensors, A wireless communication device through which the automated entity can wirelessly connect to at least one other automated entity among a plurality of automated entities, A memory, And at least one processor coupled to the memory, wherein the at least one processor is configured to execute one or more of the following instructions: Evaluate an initial individual rank based on its own communication and non-communication capabilities, and generate an individual planning strategy for reaching the convergence area under individual optimization constraints, Convey the evaluated initial rank and capability data, Based on the initial rank and capability data received from the responding automated entity, establish communication with the responding automated entity, and update the individual rank and planning strategy to reach the convergence area while maintaining connectivity with these responding automated entities under the constraints of cluster formation optimization.

19. The automated entity according to claim 18, wherein, As a scout, the automated entity determines its search step size and trajectory towards the convergence area based on the condition of maintaining connectivity with the responding automated entity and the follower(s).

20. The automated entity according to claim 18, wherein, As a scout, if the automated entity estimates that it will lose connection with its nearest follower(s) when moving according to the determined step size and trajectory, it conveys its own search step size and trajectory to the follower.

21. The automated entity according to claim 18, wherein, As a scout, the automated entity periodically sends its own calculated step size and trajectory to the follower(s).

22. The automated entity according to claim 18, wherein, As a scout, the automated entity periodically calculates the step size and trajectory data of itself and the follower(s), and sends this data to the follower(s).

23. The automated entity according to claim 18, wherein, In the absence of any response received from any other automated entity among the plurality of automated entities, the automated entity operates autonomously according to the individual planning strategy.

24. The automated entity according to claim 18, wherein, As a follower, the automated entity operates according to the step size and trajectory provided by the scout by correspondingly updating its own step size and trajectory.

25. The automated entity according to claim 18, wherein As a follower, if it is determined that it cannot follow the provided step size and trajectory due to its limited capabilities, the automated entity indicates its last position to the scout.

26. The automated entity according to claim 18, wherein, As a follower, the automated entity forwards any indication of the last position received from other followers to the scout.

27. The automated entity according to claim 18, wherein As a scout, the automated entity records any indication of the last position provided by the follower.

28. The automated entity according to claim 18, wherein, The automated entity includes robots, driverless and autonomous systems, automated guided vehicles, unmanned aerial vehicles, drones, or high-altitude mobile platforms.

29. The automated entity according to claim 18, wherein The automated entity is an aircraft having wide-area and local-area wireless communication capabilities and acting as a communication repeater or mobile base station.

30. A cluster of automated entities that are called to perform a mission, wherein, The swarm includes the automated entities according to claims 18 to 29, including combinations thereof, wherein the aerial entities are capable of supporting the ground entities in a distributed swarm formation.

31. A computer-readable medium for distributing formations of automated entities and storing computer-executable instructions, wherein, The computer-executable instructions, when executed, cause at least one processor to perform the method according to any one of claims 1 to 17.

Images