A method and apparatus for polar penguin monitoring

Abstract

The application provides a method and device for polar penguin monitoring, and relates to the technical field of ecological monitoring.The method comprises the following steps: establishing a local communication subnet and synchronizing absolute time stamps, calculating the attenuation boundary time of each node based on the current environmental absolute temperature; acquiring the absolute water depth, relative distance and motion characteristic value of each node in the local communication subnet, and constructing a cluster motion centroid vector; calculating the initial expected observation utility value of each node based on the cluster motion centroid vector, the absolute water depth and the attenuation boundary time, setting the node with the highest initial expected observation utility value in the local communication subnet as the main node for starting the recording of the camera module, and setting the remaining nodes as slave nodes; the main node calculates the real-time effective observation utility based on the spatial deviation, and calculates the utility difference time integral in combination with the real-time alternative utility value broadcast by the slave nodes; when the utility difference time integral breaks through the threshold value, the main node sends a handover signal to the slave nodes for dynamic relay recording.

Description

Technical Field

[0001] This invention relates to the field of ecological monitoring technology, and more specifically, to a method and equipment for monitoring polar penguins. Background Technology

[0002] As apex predators and keystone species in polar ecosystems, penguins' population dynamics, habitat selection, and group foraging behavior are important indicative parameters reflecting the health of polar marine ecosystems and global climate change trends. To comprehensively and sustainably obtain relevant ecological data, researchers typically employ biological loggers or remote sensing tracking devices. Current monitoring equipment primarily consists of physical structures fixed to the penguin's body surface, integrating microcontrollers, satellite positioning modules, environmental sensors, and camera modules to acquire image information and kinematic characteristics of the surrounding environment when the animals exhibit specific behaviors.

[0003] However, the unique characteristics of the polar environment present severe challenges to the underlying hardware and environmental interaction of existing animal monitoring equipment. The polar regions are characterized by extreme low temperatures, with the normal operating temperature range of equipment often dropping to minus forty degrees Celsius or below. Under these extreme low-temperature conditions, the internal polarization of conventional chemical batteries intensifies, significantly increasing the battery's internal resistance and causing a non-linear decline in its actual usable capacity. When the load power suddenly increases, such as when the high-definition camera module is activated, it can easily trigger a sharp drop in power supply voltage or even system shutdown. Furthermore, the long periods of polar night in the polar environment render energy replenishment solutions relying on solar photovoltaic panels completely ineffective within months. Since polar penguins live in wild areas without network coverage, and equipment retrieval is extremely difficult, the limited power capacity and fixed storage space of the equipment become rigid bottlenecks restricting the entire monitoring cycle.

[0004] In existing technical solutions, monitoring devices mostly operate in an isolated, single-node mode. To extend battery life, these devices often rely on a single sensor threshold as the trigger condition for activating the camera module. This static and fixed triggering mechanism exhibits an extremely high false trigger rate and redundancy in the complex polar underwater environment. Penguins' non-predatory maneuvers, evasion of predators, or group companionship behaviors underwater can easily satisfy the single threshold condition, thus ineffectively activating the camera system. The acquisition and writing of high-definition image streams are characterized by high energy consumption and high storage consumption; frequent invalid triggers will rapidly deplete the limited hardware resources of a single device. Once the battery level drops below the threshold or the storage chip is full, the device will permanently lose the ability to record crucial predatory images and even maintain basic positioning and tracking.

[0005] On the other hand, penguins are typical social animals, and their key life activities, especially underwater hunting and foraging, exhibit a high degree of spatial clustering and behavioral coordination. Existing isolated monitoring equipment can only mechanically record individual first-person perspective information and cannot perceive the status of surrounding companions. When multiple devices are deployed simultaneously in the same penguin colony, due to the lack of communication interaction and environmental collaborative assessment mechanisms, multiple devices usually concurrently start high-power recording in the same homogeneous scene, resulting in a serious waste of the overall colony's hardware resources. At the same time, when a core group foraging event of great scientific research value occurs, the device at the best observation angle may prematurely enter dormancy due to ineffective power consumption in the early stages, leading to the loss of valuable spatiotemporal data. Therefore, how to overcome the limitations of single-device battery life and storage under the harsh constraints of extreme cold and lack of public network coverage, and efficiently acquire high-quality images of group behavior, is an objective technical problem that urgently needs to be solved in the field of polar animal monitoring technology. Summary of the Invention

[0006] This invention provides a method for monitoring polar penguins, the method comprising: A local communication subnet is established and absolute timestamps are synchronized. The decay boundary time of each node is calculated based on the current absolute temperature of the environment. The absolute water depth, relative distance, and motion characteristic values ​​of each node in the local communication subnet are obtained, and the cluster motion centroid vector is constructed. Based on the cluster motion centroid vector, absolute water depth, and decay boundary time, the initial expected observation utility value of each node is calculated. The node with the highest initial expected observation utility value in the local communication subnet is set as the master node to start recording with the camera module, and the other nodes are set as slave nodes. During the recording, the master node calculates the real-time effective observation utility based on the spatial deviation and calculates the utility difference time integral by combining the real-time alternative utility values ​​broadcast by the slave nodes. When the utility difference time integral exceeds the threshold, the master node sends a handover signal to the slave nodes to perform dynamic relay recording.

[0007] The decay boundary time of each node is calculated based on the current ambient absolute temperature, specifically including: reading the node... Current remaining physical power Current remaining storage space and the current absolute temperature ; Calculate the ambient temperature penalty coefficient : in, This represents the preset empirical extreme cold attenuation coefficient. This indicates the preset critical temperature for extreme cold attenuation. Based on the current remaining physical power To obtain the effective remaining power ; Power consumption per unit time when the camera module starts recording video Storage consumption rate per unit time The battery cold degradation phenomenon is converted into a safe operating limit for the camera module, i.e., the degradation boundary time. : .

[0008] Obtain the absolute water depth, relative distance, and motion characteristic values ​​of each node within the local communication subnet, specifically including: Calculated nodes Motion eigenvalues ;in, Indicates the start time of the current sampling period. This indicates the preset integration time window duration. Represents a three-dimensional linear acceleration vector. Represents a three-dimensional angular velocity vector. This represents the preset rotation weighting coefficient. The Euclidean norm of a vector; At the same time, the parsing receives neighboring nodes. Received signal strength indicator value when broadcasting data packets Calculate the nodes With nodes relative distance between .

[0009] Constructing the cluster's motion centroid vector specifically includes: establishing a local three-dimensional coordinate system with its own physical location as the origin, and calculating adjacent nodes by combining the absolute water depth and the measured relative distance. Relative to node spatial coordinate vector ; Adjacent nodes Motion characteristics of broadcast As the corresponding spatial coordinate vector The weighting coefficients automatically shift the calculated three-dimensional coordinate vector towards the area where predation activity is most concentrated. Construct the cluster motion centroid vector representing the focal point of community cooperative behavior. ;in, Represents the current time and node Maintain the set of all neighboring nodes for communication connections. Represents a node Its own motion characteristic values.

[0010] Before calculating the initial expected observation utility value for each node, the microcontroller needs to evaluate the directional observation value and optical imaging conditions of the camera module, specifically including: Time integration of three-axis acceleration data to obtain nodes At the present moment Three-dimensional velocity vector Combined with the cluster motion centroid vector and its Euclidean norm Extract the cosine value of the spatial angle between the current swimming direction and the direction of the centroid to obtain the observation angle fraction. ; Combined nodes absolute water depth By penalizing optical imaging failure nodes caused by the polar deep-water environment, an optical feasibility weight is obtained. : in, This represents the light attenuation coefficient of polar seawater.

[0011] The steps of calculating the initial expected observation utility value for each node and setting the corresponding node as a master node or a slave node specifically include: Node At the decay boundary time Based on the current observation perspective, the score Weighting of optical feasibility Generate nodes Initial expected observed utility value : Generate nodes Initial expected observed utility value ; Each microcontroller broadcasts the initial expected observation utility value to the local communication subnet via the wireless communication module to perform numerical comparison. The node that confirms that its value is the highest connects the camera module to power on and is set as the master node, while the other nodes keep the camera module in sleep mode and are set as slave nodes.

[0012] In the steps of calculating the real-time effective observation utility based on spatial deviation, the master node first performs a spatial deviation assessment, which specifically includes: during the recording period of the camera module, the master node continuously extracts its own time at time... Three-dimensional velocity vector It also receives broadcast data to update the time in real time. Cluster motion centroid vector The master node is obtained by extracting the cosine value of the spatial angle between its current swimming direction and the direction of the community's predation core. At any moment Spatial deviation : in, This represents the Euclidean norm of a vector.

[0013] The steps of the master node calculating the real-time effective observation utility based on spatial deviation further include: [the steps of] spatial deviation... As an exponential penalty term, the initial expected observed utility value Dynamic reduction is performed to obtain the master node. At any moment Real-time effective observation utility in, This represents the preset divergence penalty weight constant.

[0014] The steps for dynamic relay recording, which combine the real-time candidate utility values ​​broadcast from the nodes to calculate the time integral of the utility difference, specifically include: The master node listens to all candidate slave nodes that are in a standby state. At the moment of integration variable Real-time alternative utility value of broadcasting To filter out false handover requests caused by short-term attitude fluctuations, a preset anti-shake hysteresis time window is used. Within, for exceeding the handover hysteresis tolerance threshold Integrating the utility difference over time yields the result at time [time value missing]. To candidate slave nodes Utility difference time integral : in, Indicates the time of the master node. Real-time and effective observation utility; When the utility difference is integrated over time When the accumulated value exceeds the preset execution threshold, the master node sends a handover signal to the candidate slave node. The candidate slave node becomes the new master node and starts recording, while the original master node becomes a slave node and enters low-power standby mode.

[0015] The present invention also provides a device for monitoring polar penguins, including an installation mechanism and a monitoring body that are spliced ​​together, and a reversible disassembly device is provided between the monitoring body and the installation mechanism; the monitoring body includes a shell, a microcontroller encapsulated inside the shell, and an energy supply module, a data transmission module, a satellite positioning module, a behavior monitoring module, a water depth monitoring module, a water entry detection module and a camera module that are electrically connected to the microcontroller; The microcontroller is configured as follows: It establishes a local communication subnet and synchronizes absolute timestamps via the data transmission module, calculating the decay boundary time based on the current ambient absolute temperature; after the water entry detection module is activated, it acquires absolute water depth, relative distance, and motion characteristic values, fusing them to construct a cluster motion centroid vector; based on the cluster motion centroid vector, absolute water depth, and decay boundary time, it calculates the initial expected observation utility value, setting the node with the highest utility value within the local communication subnet as the master node to start recording via the camera module, and the others as slave nodes; during recording, the master node calculates the real-time effective observation utility based on spatial deviation, and calculates the utility difference time integral by combining the real-time alternative utility values ​​broadcast by the slave nodes; when the integral exceeds the threshold, it sends a handover signal via the data transmission module to perform dynamic relay recording.

[0016] This invention provides a method and device for monitoring polar penguins. Utilizing the gregarious behavior of penguins before entering the water, this application pre-synchronizes absolute time within a local area network, eliminating the time axis misalignment problem caused by satellite timing signal loss in the polar underwater environment. Regarding the underwater event triggering mechanism, this application overcomes the technical bottleneck of traditional monitoring equipment that relies on a single depth or acceleration threshold, easily generating massive amounts of redundant data. By integrating the absolute water depth of individuals, three-dimensional maneuvering states, and relative spatial topological distances within the local area network, the cluster motion centroid, representing the focus of coordinated predation, can be dynamically calculated without relying on external absolute coordinates. Taking into full account the rapid acceleration and high-frequency angular tumbling characteristics of penguins during predation, ineffective high-energy-consuming actions during routine cruising or simply avoiding predators are filtered out, ensuring that the system is only partially activated when the core group hunting behavior actually occurs. This significantly reduces the false trigger rate in non-target scenarios and effectively avoids the unnecessary consumption of the device's extremely limited power and storage space on invalid blank screens.

[0017] Regarding resource scheduling within the device cluster, this application employs a master-slave node dynamic allocation strategy based on utility evaluation. Due to the severe light index decay and ice cover obstruction in the polar underwater environment, this application considers the physical short distance of an individual device from the predation core area when deciding on high-definition camera nodes, while also incorporating deep-water optical feasibility, spatial dynamic approach course, and hardware safety time constraints. Through cross-constraints, the system can adaptively select the optimal individual node that combines sufficient natural ambient light, a forward approach perspective, and a sufficiently long continuous recording time. This avoids the waste of overall cluster hardware resources caused by concurrent recording by multiple devices in homogeneous scenarios, achieving globally optimized energy and storage configuration.

[0018] Regarding the process control of continuous recording, this application evaluates the highly intermittent and unpredictable nature of penguin underwater foraging maneuvers. By introducing a dynamic evaluation of spatial deviation and a hysteresis integral verification mechanism, the system can monitor the perspective changes of the main segment in real time. Once it is confirmed that the perspective of the current recording node has substantially deviated from the core of the foraging process and a better alternative exists within the communication subnet, the system smoothly transfers the camera task between different individuals via low-power local area network commands. This dynamic relay mechanism breaks the rigid physical limitation of a single device having a fixed first-person perspective, avoids ineffective high-frequency oscillations between different nodes, and transforms the originally isolated individual recording into a relay tracking of the entire group's foraging cycle. Ultimately, after recovering the equipment, researchers can integrate and output a high-definition, continuous image that is unaffected by individual disengagement maneuvers and continuously anchored to the core foraging area, providing high-quality panoramic behavioral data for polar marine ecological research. Attached Figure Description

[0019] Figure 1 This is a flowchart of the invention for monitoring polar penguins; Figure 2 This is a comparison chart of the received signal strength fluctuation curves of each node and the power consumption under the collaborative relay mechanism when the monitoring equipment is in a polar underwater communication environment. Figure 3 The evaluation curve of the real-time effective observation utility of multiple nodes in a local communication subnet and the timing diagram of the dynamic relay of the master node state machine are shown in the figure. Figure 4 This is a graph showing the dynamic evolution of ambient temperature, system terminal voltage, effective remaining power, and decay boundary time of the monitoring equipment of this invention in polar regions. Detailed Implementation

[0020] This embodiment provides a device for monitoring polar penguins. The main scenario for this invention is the polar wild environment characterized by extreme low temperatures, deep-sea high pressure, and the absence of public wireless network coverage. The device is mounted on the penguin and is used to collect multi-dimensional spatial topological features, kinematic states, and high-resolution image data during penguin colonies' terrestrial gatherings and underwater foraging.

[0021] The device mainly consists of an interconnected mounting mechanism and a monitoring unit. A reversible disassembly device is provided between the monitoring unit and the mounting mechanism. Preferably, the bottom surface of the mounting mechanism is bonded to the surface of the penguin's back feathers using low-temperature resistant structural adhesive; the reversible disassembly device employs a mechanical clamping and quick-release locking structure to allow for repeated installation and removal of the monitoring unit during monitoring tasks at different stages of its life cycle, thereby avoiding damage to the inherent waterproof and heat-insulating properties of the penguin feathers from repeated adhesive application.

[0022] The monitoring entity is the execution end of this embodiment, which includes a housing and a microcontroller, a power supply module, a data transmission module, a satellite positioning module, a behavior monitoring module, a water depth monitoring module, a water entry detection module, and a camera module encapsulated inside the housing.

[0023] The outer shell is integrally molded from a high-strength, lightweight polymer material resistant to low temperatures. To resist high-pressure leakage in the polar underwater environment, a defect-free encapsulation layer is cured inside the shell. Specifically, this defect-free encapsulation layer is prepared using a two-stage vacuum potting process.

[0024] The microcontroller serves as the central processing and collaborative control core of the monitoring system. Preferably, the microcontroller uses a low-power main control chip, with the GigaDevice GD32F470VIH6 chip being a preferred choice. This microcontroller is electrically connected to various peripheral hardware modules via an internal communication bus, and is responsible for parsing sensor data and executing network control logic and task offloading instructions at the node level.

[0025] The energy supply module includes a lithium battery, a solar charging panel disposed on the upper surface of the outer casing, and a charging probe disposed on the outer side of the outer casing. The solar charging panel is used to absorb solar energy to replenish the system during polar days; the charging probe preferably adopts a Pogo Pin charging interface for connecting an external charging device after the equipment is physically recycled. The lithium battery serves as the main power source, and its instantaneous output voltage and remaining usable capacity percentage are monitored by a microcontroller.

[0026] The data transmission module includes a wired data transmission interface and a wireless Bluetooth communication module. The wired data transmission interface is used for offline massive data export after the device is detached and recycled; the Bluetooth communication module is configured not only to respond to connection requests from the host computer, but also to periodically send and listen for low-power broadcast packets when the device is in operation. The microcontroller collects signals transmitted by neighboring nodes through the Bluetooth communication module and analyzes the received signal strength indication data to achieve node discovery and relative distance perception in a local environment.

[0027] The satellite positioning module employs a high-precision multi-mode positioning chip, preferably the MC60 module developed by Shanghai Quectel. This module supports multiple satellite navigation systems including BeiDou, GPS, GLONASS, and QZSS, and is used to acquire absolute azimuth, longitude, latitude, and altitude data when penguins are floating on water to breathe or engaging in land activities and satellite signals are not blocked, and then writes the positioning data into the memory.

[0028] The behavior monitoring module, water depth monitoring module, and water entry detection module work together to form a spatial and state perception unit. The behavior monitoring module preferably uses the STMicroelectronics LSM6DSV32X module, which integrates a three-axis accelerometer and gyroscope to collect real-time motion vectors and intensity characteristics of the device in three-dimensional space. The water depth monitoring module uses a high-pressure water depth sensor to collect real-time absolute depth characteristics along the Z-axis in underwater environments where satellite positioning signals are lost. The water entry detection module includes a positive probe IN+ and a negative probe IN-, both exposed through the outer casing and connected to the internal main control operational amplifier circuit (e.g., using a TLV521DCKR chip). The output pin OUT of this water entry detection module is configured as follows: a high level when the probe is in air and non-conductive, and a low level when the probe is immersed in water and conductive. The microcontroller records the state transition timestamp based on the transition edge of this level.

[0029] The camera module is housed within a high-transparency window at the front of the monitoring unit, and preferably employs the OV5640 high-performance CMOS image sensor designed and manufactured by OMNIVISION. This camera module features a five-megapixel resolution, supports autofocus control, and receives start / stop control signals from the microcontroller via the IIC bus, thereby switching between sleep / standby mode and video recording mode.

[0030] Next, this embodiment provides a method for monitoring polar penguins based on the hardware architecture.

[0031] S1: Reference synchronization and attenuation modeling: Before initiating their group swimming behavior, polar penguins typically congregate densely on ice or unobstructed water surfaces. During this stage, the environment is unobstructed by high-density seawater, ensuring stable radio frequency signal transmission. The microcontrollers of each monitoring unit activate their satellite positioning modules to receive satellite timing signals in this unobstructed environment and obtain the current absolute timestamp. The microcontroller controls the wireless Bluetooth communication module to enable broadcast and scan modes. Monitoring entities within each other's communication range complete a handshake by intercepting each other's broadcast signals and establish a point-to-point local communication subnet based on the Received Signal Strength Indication (RSSI).

[0032] During the establishment of the local communication subnet, the microcontroller uses the node within the network that acquires the highest quality satellite timing signal as the timing reference node. This timing reference node uses an absolute timestamp. The data is transmitted via a wireless Bluetooth communication module to all nodes within the local communication subnet. Each node receives the absolute timestamp. Then, the internal real-time clock (RTC) is forced to calibrate and align.

[0033] After completing time synchronization, the microcontroller needs to assess its own hardware sustainability. This application employs a linear penalty algorithm based on a temperature threshold to convert the battery's cold degradation into a direct time limit for the camera module to operate.

[0034] Nodes are read in real time using built-in sensors. Current remaining physical power Current remaining storage space and the current absolute temperature Compute nodes Ambient temperature penalty coefficient , in, This indicates the node identifier sequence number of the monitoring entity within the local communication subnet; Represents a node The calculated ambient temperature penalty coefficient; This represents the preset empirical extreme cold attenuation coefficient, used to indicate the rate of decrease in battery discharge efficiency caused by a unit temperature drop; This indicates the preset critical temperature for extreme cold attenuation. Represents a node The current absolute temperature of the environment was collected; This represents the function that takes the maximum value.

[0035] Preferably, for the low-temperature resistant modified lithium-ion battery pack selected for polar monitoring equipment, the empirical extreme cold degradation coefficient is... The value is set to Meanwhile, the preset critical temperature for extreme cold attenuation Set as The freezing point of polar seawater is usually between [insert value here]. Around, when the ambient temperature drops to At the following time points, the viscosity of the electrolyte inside the battery begins to increase significantly, and the ionic conductivity shows a substantial decline. Based on the discharge decay curves of low-temperature resistant lithium batteries under extreme cold testing environments, in... to Within the core operating temperature zone, every time the temperature drops The actual usable capacity of a battery under a constant high current load is approximately reduced by an average linear rate. .

[0036] Based on this, compute nodes Effective remaining power The formula is as follows: in, Represents a node At the current absolute temperature Effective remaining power under constraints; Represents a node The current remaining physical battery power is read directly. Combined with camera module parameters, the node's power is calculated. decay boundary time : in, Represents a node The decay boundary time, which is the safe limit for the node to continuously record video; This represents a function that takes the minimum value. This indicates the power consumption per unit time when the camera module starts recording video; This indicates the storage consumption rate per unit time when the camera module starts recording video; Represents a node The current remaining storage space.

[0037] When polar penguins enter the water, the ambient temperature drops suddenly to freezing point or below. At this time, relying solely on reading the current remaining physical battery power will not be sufficient. The decision-making process involves addressing the issue that activating the high-power camera module would cause the battery to shut down due to a sharp voltage drop caused by low-temperature polarization. This invention utilizes the critical temperature for battery degradation in extreme cold conditions. With the maximum value function, when the water temperature is above the critical temperature, the penalty term is zero, avoiding redundant calculations; when entering extremely cold waters, the chemical polarization effect can be quickly converted into decay boundary time using only subtraction and multiplication logic in the microcontroller. By obtaining This yielded the judgment constraints for the collaboration and task execution of each node, which not only ensured the security of the attenuation assessment but also greatly saved the computing power of the edge devices.

[0038] Existing wildlife monitoring equipment typically relies on single-dimensional sensor data to trigger video recording, such as directly recording when a set absolute water depth is exceeded or a certain instantaneous acceleration threshold is breached. However, penguins in polar waters engage in various behaviors beyond hunting, including regular patrolling, companionship, and maneuvering to evade predators. These behaviors generate significant changes in kinematic parameters. Using independent thresholds for detection could easily lead to the camera module being ineffectively activated in non-hunting scenarios, rapidly depleting the device's extremely limited remaining battery power and storage space.

[0039] Since underwater hunting by penguins is a highly coordinated group behavior, in order to overcome the problem of false triggering caused by threshold determination of a single sensor, this embodiment provides a feature space construction method based on multi-source sensor fusion. The following describes step S2 of the method: feature space construction based on multi-source sensor fusion.

[0040] When the penguin submerges into the seawater, the two probes of the water entry detection module of the monitoring device become conductive due to the seawater medium. The microcontroller captures this level transition signal and determines that the device has entered an underwater operating environment. Because the polar seawater environment has a strong absorption and attenuation effect on microwave signals, continuing to maintain conventional satellite positioning would result in ineffective and high power consumption. Therefore, the microcontroller immediately switches the satellite positioning module to a low-frequency sleep mode, retaining only its wake-up capability when surfacing for air. Simultaneously, the microcontroller controls the behavior monitoring module and the depth monitoring module to enter a high-frequency data sampling state, and maintains wireless Bluetooth broadcasting and listening within the local communication subnet to compensate for the underwater spatial perception blind spots caused by the lack of absolute positioning.

[0041] During underwater high-frequency sensing, the microcontroller calculates the motion characteristic value of the current node within a preset time window by fusing the three-axis acceleration data output by the behavior monitoring module and the gyroscope data. The purpose of this step is to extract the high-energy-consuming movement characteristics of penguins when hunting fish, such as rapid acceleration and turning.

[0042] in, Represents a node Motion characteristic values ​​obtained through integration; Indicates the start time of the current sampling period; This indicates the preset integration time window duration; Represents a node At any moment The three-dimensional linear acceleration vector is read; Represents a node At any moment The three-dimensional angular velocity vector read; This represents the preset rotation weighting coefficient; This represents the Euclidean norm of a vector.

[0043] During the penguin's hunting process, chasing schools of fish is accompanied by violent rolling and changing direction. By introducing the norm integral of the three-dimensional angular velocity and the rotation weight coefficient, this invention can comprehensively map the straight-line rush and angular changes into a scalar that reflects the current level of individual behavioral activity, thereby filtering out the penguin's smooth swimming data during normal cruising.

[0044] While acquiring its own motion characteristics, the monitoring subject needs to further perceive the spatial distribution of the community it inhabits. Node Receive adjacent nodes within the local communication subnet The broadcast data packet sent contains a node. Absolute water depth collected and its calculated kinematic characteristic values When parsing the received broadcast data packet, obtain the corresponding received signal strength indicator. And calculate the relative distance. : In the above formula, Indicates being in a node The sequence number of adjacent nodes within the communication range; Represents a node With nodes The relative physical distance between them; Represents a node Received node The received signal strength indicator value when the signal is received; This represents the environmental reference signal strength constant, which is the reference signal strength value when the two nodes are one meter apart; The path loss attenuation index represents the path loss attenuation index in polar marine environments.

[0045] Preset underwater environment reference signal strength Preferred setting is Path loss attenuation index in polar marine environments Preferred setting is Microwave signals in the 2.4 GHz band exhibit extremely high dielectric loss when propagating underwater. In typical open-line-of-sight environments, the path loss attenuation index... The value is usually taken from Around; while in complex indoor multipath environments, the value is approximately Polar seawater, due to its high salinity, has a much higher electrical conductivity than freshwater, resulting in a strong absorption effect on the high-frequency components of electromagnetic waves. (Preset benchmark) This indicates the initial radiation attenuation after the equipment casing is enclosed by water in a polar ice-water mixture. This is a typical empirical constant derived from channel testing in shallow, high-salinity polar seawater environments. Using this relatively high attenuation index can correct for the non-linear, sharp drop in Bluetooth signal strength with distance in seawater, thus ensuring the accuracy of the relative physical distance between nodes calculated by the microcontroller. There will be no severe scale divergence.

[0046] After obtaining the above relative physical distances, the node The microcontroller establishes a local three-dimensional coordinate system with its own physical location as the origin, and combines this with its own absolute water depth. Absolute water depth of adjacent nodes and the measured relative distance Calculate the nodes Relative to node spatial coordinate vector .

[0047] Based on depth parameters, spatial distribution characteristics, and the motion state of each node, a cluster motion centroid vector is constructed to characterize the focal point of community cooperative behavior. in, Represented by node The centroid vector of the cluster motion calculated with the origin of the coordinate system as the reference point; Represents the current time and node Maintain the set of all neighboring nodes with which the communication connection is established; Indicates adjacent nodes Motion characteristic values ​​of broadcasting; Represents a node Relative to node Spatial coordinate vector; Represents a node Its own kinematic characteristic values.

[0048] In actual polar underwater swarm foraging, penguin colonies exhibit highly coordinated foraging behavior. The group uses coordinated movements to drive prey and gather it into densely populated areas easy to capture. During this process, individuals directly involved in the pursuit and capture typically exhibit significantly higher kinematic characteristic values ​​due to frequent rapid accelerations and sharp turns, while individuals in peripheral coordinated or accompanying positions show relatively low-intensity, steady swimming behavior. This invention calculates the swarm motion centroid vector by using the kinematic characteristic values ​​of neighboring nodes as weighting coefficients for their corresponding spatial position vectors. This allows the calculated three-dimensional coordinate vector to automatically shift towards the area of ​​highest concentration and most intense foraging activity within the group. After obtaining this swarm motion centroid vector, each node can determine its deviation from the core foraging area in the entire three-dimensional ocean space, thus fusing the independent and scattered sensor data of each node into a unified spatial topological representation of swarm behavior.

[0049] After obtaining the centroid vector of the cluster motion in the core area of ​​the community's predation through step S2, the system faces the challenge of determining the optimal recording node, i.e., the master node, within the local communication subnet.

[0050] Since the monitoring device in this application is a tracking device fixed to the penguin's back, its camera module has a relatively fixed field of view, only recording the view directly in front of the penguin's torso. If only the shortest spatial distance is used as the allocation criterion, it is easy to select nodes that, although close to the core predation area, are moving away from the core area, such as nodes that have just completed a dive and are turning away from a school of fish, resulting in recording blank water areas of little value. In addition, polar seawater has strong absorption and scattering of natural light. As the water depth increases, the ambient illuminance decreases exponentially. Even if the viewing angle is correct, nodes in deep water may still produce low-quality images due to the lack of basic imaging illuminance.

[0051] To address the high-power, ineffective recording issues caused by the limitations of a fixed viewing angle and optical attenuation in deep polar water, this embodiment employs a predictive master node allocation method based on a utility function. The following details S3: Predictive Master Node Allocation Based on a Utility Function.

[0052] The microcontroller first needs to determine whether the camera module is facing a high-value target. Since the penguin's main thrust direction underwater is consistent with the height of its body's longitudinal axis, the three-dimensional velocity vector of the current node is obtained by time integration of the three-axis acceleration data output by the behavior monitoring module, thereby characterizing the instantaneous spatial orientation of the camera module.

[0053] Based on the obtained cluster motion centroid vector, the node is calculated. observation perspective score : Represents a node The calculated observation angle score; represents the Euclidean norm of the centroid vector of the cluster's motion, and represents the absolute spatial distance between the node and the predation core region; Represents a node At the present moment The three-dimensional velocity vector; The exponential decay term quantifies the impact of distance; the farther from the centroid, the lower the base score. The second term extracts the cosine of the spatial angle between the current swimming direction and the centroid direction through vector dot product. When the penguin is accelerating towards the core predation area, the angle between the two vectors approaches zero, and the cosine value is at its maximum in the positive direction. At this point, not only is the distance close, but the camera module is also perfectly aligned with the school of fish, resulting in a higher observation angle score. Gain a positive gain; if the penguin is swimming away from the core area, the cosine value is negative, even if it is currently... Even if the value is very small, its observation score will be forcibly weakened, thereby eliminating invalid nodes that are close at hand but have incorrect perspectives.

[0054] After confirming that the nodal perspective has value for directional observation, it is necessary to further filter out nodals that fail in optical imaging due to the deep-water polar environment. Penguins hunt by diving to great depths, and the strong absorption of visible light by seawater leads to a decrease in underwater illumination.

[0055] The microcontroller calls the absolute water depth data output by the water depth monitoring module and calculates the node based on the underwater light intensity decay law. Optical feasibility weight : The light attenuation coefficient of polar seawater, preferably, the light attenuation coefficient The empirical range is to In this embodiment, the microcontroller's non-volatile memory presets and defaults to calling fixed empirical intermediate values. Perform the calculation.

[0056] When two nodes are located above and below the predation core area respectively, and are approximately equidistant from the center of mass, the node located in the deep water area will be more likely to be affected. The optical feasibility weight is relatively large. This will be penalized. The system will naturally tend to select nodes in relatively shallow water areas with better natural lighting as video sources, avoiding the black screen and useless data caused by blindly activating deep-water node camera modules, thus ensuring the resolution of the image data.

[0057] After obtaining the optical feasibility of the observation viewpoint score and depth, the aforementioned environmental and behavioral parameters are analyzed in conjunction with the hardware persistence status. This yields the node... Initial expected observed utility value : Regardless of the optical depth conditions of the node How superior, or sprint perspective How precise! As soon as its battery is about to face voltage collapse due to extremely cold temperatures, that is... Approaching zero, its expected observational utility value for accidents is... All such attempts will be rejected outright. This ensures that the nodes selected by the system must have the basic power capacity to record a complete and valid predation video.

[0058] The expected observation utility value, calculated using a multiplication operation, represents the theoretically longest shooting time for a node under perfect conditions (sufficient lighting, direct view of the target). The lifespan is discounted proportionally based on its current actual unfavorable position and viewing angle deviation. The final output time unit indicates how long a node can continuously output high-resolution and aligned high-quality predation footage if the camera module is enabled.

[0059] Generate initial expected observation utility values Subsequently, each microcontroller broadcasts this value to the local communication subnet via a wireless Bluetooth communication module. All online nodes receive the utility values ​​of their neighboring nodes within a preset synchronization time window and perform value sorting and comparison within the microcontroller. When a node... Its own was confirmed through comparison. When the value is the highest within the current local communication subnet, the microcontroller actively sets its system state to master node. The master node connects the main power supply circuit of the camera module, starts high-definition video recording, and saves the video stream containing the system timestamp to memory. Conversely, when the value is lower than the highest within the local communication subnet, the system state is confirmed through comparison. For nodes other than those with the highest value, the microcontroller actively sets their status to slave node. The slave nodes maintain the camera module's sleep and power-off state to conserve power, but continue to monitor the wireless Bluetooth communication module and sample from various sensors, continuously performing utility evaluation for the next cycle in the background.

[0060] After establishing the master node through predictive allocation in step S3 and initiating recording by the camera module, the system enters a continuous data acquisition phase. Existing monitoring equipment, upon triggering recording, continues recording until the set maximum recording time ends or the battery is physically depleted. However, penguin underwater hunting maneuvers in polar regions are highly intermittent and unpredictable. In real-world group hunting, once the master node penguin successfully catches prey, it typically immediately performs maneuvers such as sudden braking, swallowing, or rapid surfacing for air. These behaviors cause the individual master node's trajectory to deviate from the main penguin group that continues chasing the fish, resulting in a significant deviation from the group's center of mass. In this state, allowing the master node to continue recording without dynamic feedback intervention not only causes it to lose its perspective and record a large amount of worthless blank water footage but also wastes the limited available power of the monitoring equipment in extreme cold weather.

[0061] The following section elaborates on S4: Continuity verification and dynamic relay based on actual-prediction deviation.

[0062] The master node performs spatial deviation assessment based on spatial trajectory deviation. During camera module recording, the master node continuously extracts its own 3D velocity vector and receives sensor broadcast data from slave nodes within the local communication subnet to update the cluster's motion centroid vector in real time. To capture the penguin's turning-around action after catching prey, the microcontroller calculates the master node's... At the present moment Spatial deviation : The fractional term extracts the cosine value of the spatial angle between the current swimming direction of the master node and the direction of the core predation of the fish in the school. When the master node closely follows the main group to chase the fish school, the angle approaches zero degrees, the cosine value approaches one, and the spatial deviation is... The inhibition converges to zero; when a penguin completes a feeding maneuver and then brakes suddenly, swallows, or rapidly rises to breathe, the angle widens, the cosine value turns negative, and the spatial deviation increases. It rapidly approaches the maximum value, thus quantifying the degree of detachment from the animal's perspective.

[0063] Spatial deviation represents the deterioration of the current viewpoint. The microcontroller uses this deviation as an exponential penalty term to dynamically reduce the expected observation utility value generated in step S3, and calculates the real-time effective observation utility of the master node. : in, Indicates the master node At any moment Real-time and effective observation utility; This represents the preset divergence penalty weight constant.

[0064] Once the master node undergoes a violent divergence maneuver, The value will drop, reflecting the diminishing value of continuing to record.

[0065] The master node continuously calculates its own real-time effective observation utility. Simultaneously, the system uses a wireless Bluetooth communication module to monitor the real-time alternative utility values ​​broadcast by each slave node in standby mode within the local communication subnet at fixed intervals. When penguins are weaving and changing direction underwater, they may only be briefly avoiding their companions. If they immediately switch recording tasks once they find that the node's utility has surpassed theirs, it can easily cause switching oscillations, resulting in high-frequency and disordered switching of camera tasks between different individuals. This leads to a large number of fragmented waste frames and consumes additional communication power.

[0066] For candidate slave nodes Calculate the time integral of the utility difference : in, Indicates the current moment The calculated candidate slave nodes Utility difference time integral; This represents the sequence number of a candidate slave node within a local communication subnet; This indicates the preset anti-shake hysteresis time window length; Indicates candidate slave node At the moment of integration variable The real-time alternative utility value of broadcasting; Indicates the time of the master node. Real-time and effective observation utility; This indicates the preset handover hysteresis tolerance threshold.

[0067] Utility difference time integral requirement candidate slave node The utility value must not only exceed the master node, but the extent of the exceedance must also be greater than the hysteresis tolerance threshold. And simultaneously throughout the entire time window Maintaining this absolute advantage is crucial. Only a sustained and significant overtaking can effectively accumulate the time integral of the utility difference, thereby reliably filtering out false handover requests caused by short-term attitude fluctuations.

[0068] When the utility difference is integrated over time When the accumulated value exceeds the preset execution threshold, the master node... The master node determines that its own perspective has completely failed and that a better alternative exists within the subnet. At this point, the master node... Send the candidate slave node via wireless Bluetooth communication module Send with absolute timestamp The handover signal. From the node Upon receiving the command, its microcontroller connects the main power supply circuit of the camera module, becoming the new master node and immediately starting recording. (Original master node) Upon receiving from the slave node After receiving the response frame, the device completely cuts off power to its own camera module, degrading to a low-power standby state. This dynamic relay mechanism overcomes the limitations of a single device's fixed perspective. By utilizing redundant device nodes within the local communication subnet to construct a dynamic camera matrix, it not only promptly cuts off invalid recordings from the master node to conserve remaining power, but also, through unified time-base alignment, allows the device to be retrieved later and, in the host computer system, precisely stitch together fragmented video streams recorded by different nodes on different timelines into a single high-definition continuous panoramic image that is always anchored to the core area of ​​the penguin colony's foraging.

[0069] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0070] In this specification, the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the descriptions of the embodiments described later are relatively simple, and relevant parts can be referred to the descriptions of the foregoing embodiments.

[0071] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring polar penguins, characterized in that, The method includes: Establish a local communication subnet and synchronize absolute timestamps, and calculate the decay boundary time of each node based on the current ambient absolute temperature; Obtain the absolute water depth, relative distance, and motion characteristic values ​​of each node within the local communication subnet, and construct the cluster motion centroid vector; Based on the cluster motion centroid vector, absolute water depth, and decay boundary time, the initial expected observation utility value of each node is calculated. The node with the highest initial expected observation utility value in the local communication subnet is set as the master node that starts recording with the camera module, and the other nodes are set as slave nodes. During recording, the master node calculates the real-time effective observation utility based on the spatial deviation, and calculates the utility difference time integral by combining the real-time alternative utility values ​​broadcast by the slave node. When the utility difference time integral exceeds the threshold, the master node sends a handover signal to the slave node to perform dynamic relay recording. The decay boundary time of each node is calculated based on the current ambient absolute temperature, specifically including: reading the node... Current remaining physical power Current remaining storage space and the current absolute temperature ; Calculate the ambient temperature penalty coefficient : ; in, This represents the preset empirical extreme cold attenuation coefficient. This indicates the preset critical temperature for extreme cold attenuation. Based on the current remaining physical power To obtain the effective remaining power ; ; Power consumption per unit time when the camera module starts recording video Storage consumption rate per unit time The battery cold degradation phenomenon is converted into a safe operating limit for the camera module, i.e., the degradation boundary time. : ； Obtain the absolute water depth, relative distance, and motion characteristic values ​​of each node within the local communication subnet, specifically including: ; Calculated nodes Motion eigenvalues ;in, Indicates the start time of the current sampling period. This indicates the preset integration time window duration. Represents a three-dimensional linear acceleration vector. Represents a three-dimensional angular velocity vector. This represents the preset rotation weighting coefficient. The Euclidean norm of a vector; At the same time, the parsing receives neighboring nodes. Received signal strength indicator value when broadcasting data packets Calculate the nodes With nodes relative distance between ; Constructing the cluster's motion centroid vector specifically includes: establishing a local three-dimensional coordinate system with its own physical location as the origin, and calculating adjacent nodes by combining the absolute water depth and the measured relative distance. Relative to node spatial coordinate vector ; Adjacent nodes Motion characteristics of broadcast As the corresponding spatial coordinate vector The weighting coefficients automatically shift the calculated three-dimensional coordinate vector towards the area where predation activity is most concentrated. Construct the cluster motion centroid vector representing the focal point of community cooperative behavior. ;in, Represents the current time and node Maintain the set of all neighboring nodes for communication connections. Represents a node Its own motion characteristic values; Before calculating the initial expected observation utility value for each node, the microcontroller needs to evaluate the directional observation value and optical imaging conditions of the camera module, specifically including: Time integration of three-axis acceleration data to obtain nodes At the present moment Three-dimensional velocity vector Combined with the cluster motion centroid vector and its Euclidean norm Extract the cosine value of the spatial angle between the current swimming direction and the direction of the centroid to obtain the observation angle fraction. ; ; Combined nodes absolute water depth By penalizing optical imaging failure nodes caused by the polar deep-water environment, an optical feasibility weight is obtained. : ; in, This represents the light attenuation coefficient of polar seawater; The steps of calculating the initial expected observation utility value for each node and setting the corresponding node as a master node or a slave node specifically include: Node At the decay boundary time Based on the current observation perspective, the score Weighting of optical feasibility Generate nodes Initial expected observed utility value : Generate nodes Initial expected observed utility value ; Each microcontroller broadcasts the initial expected observation utility value to the local communication subnet via the wireless communication module to perform numerical comparison. The node that confirms that its value is the highest will turn on the power supply to the camera module and be set as the master node, while the other nodes will keep the camera module in sleep mode and be set as slave nodes. In the steps of calculating the real-time effective observation utility based on spatial deviation, the master node first performs a spatial deviation assessment, which specifically includes: during the recording period of the camera module, the master node continuously extracts its own time at time... Three-dimensional velocity vector It also receives broadcast data to update the time in real time. Cluster motion centroid vector The master node is obtained by extracting the cosine value of the spatial angle between its current swimming direction and the direction of the community's predation core. At any moment Spatial deviation : ; in, The Euclidean norm of a vector; The steps of the master node calculating the real-time effective observation utility based on spatial deviation further include: [the steps of] spatial deviation... As an exponential penalty term, the initial expected observed utility value Dynamic reduction is performed to obtain the master node. At any moment Real-time effective observation utility ; in, This represents the preset divergence penalty weight constant; The steps for dynamic relay recording, which combine the real-time candidate utility values ​​broadcast from the nodes to calculate the time integral of the utility difference, specifically include: The master node listens to all candidate slave nodes that are in a standby state. At the moment of integration variable Real-time alternative utility value of broadcasting To filter out false handover requests caused by short-term attitude fluctuations, a preset anti-shake hysteresis time window is used. Within, for exceeding the handover hysteresis tolerance threshold Integrating the utility difference over time yields the result at time [time value missing]. To candidate slave nodes Utility difference time integral : ; in, Indicates the time of the master node. Real-time and effective observation utility; When the utility difference is integrated over time When the accumulated value exceeds the preset execution threshold, the master node sends a handover signal to the candidate slave node. The candidate slave node becomes the new master node and starts recording, while the original master node becomes a slave node and enters low-power standby mode.

2. A device for monitoring polar penguins, used to perform the method for monitoring polar penguins as described in claim 1, characterized in that, The system comprises an interconnected installation mechanism and a monitoring unit, with a reversible disassembly device between the monitoring unit and the installation mechanism. The monitoring unit includes a housing, a microcontroller encapsulated within the housing, and an energy supply module, a data transmission module, a satellite positioning module, a behavior monitoring module, a water depth monitoring module, a water ingress detection module, and a camera module electrically connected to the microcontroller. The microcontroller is configured to establish a local communication subnet and synchronize absolute timestamps through the data transmission module, and calculate the decay boundary time based on the current ambient absolute temperature. After the water ingress detection module is activated, it acquires absolute water depth, relative distance, and motion characteristic values, and fuses them to construct a cluster motion centroid vector. The initial expected observation utility value is calculated based on the cluster motion centroid vector, absolute water depth, and attenuation boundary time. The node with the highest utility value in the local communication subnet is set as the master node to start recording with the camera module, and the rest are set as slave nodes. During the recording, the master node calculates the real-time effective observation utility based on the spatial deviation, and calculates the utility difference time integral by combining the real-time alternative utility values ​​broadcast by the slave nodes. When the integral exceeds the threshold, a handover signal is sent through the data transmission module to carry out dynamic relay recording.

Images