Alert system for use during a solo hunt or a group hunt

Abstract

The present invention relates to an alert system for use during a solo hunt or a group hunt, the system comprising at least two devices, and a remote server. The system can verify the occurrence of a firearm discharge and notify relevant users of the system's devices of such an event.

Description

Technical field of the invention

[0001] The present invention relates to the field of hunting equipment.Background of the invention

[0002] When hunters with rifles and shotguns are hunting in a dense hunting ground, such as a forest, one significant problem they face is the inability to see each other due to the thick foliage and uneven terrain. This lack of visibility makes it extremely difficult to determine the origin of any gunshot they hear. When a shot is fired, the sound can reverberate and echo through the forest, making it challenging to pinpoint its exact source. Additionally, without visual contact, hunters cannot identify who among them discharged their weapon.

[0003] This creates a potentially dangerous situation where hunters might unknowingly move into the line of fire or mistakenly assume the direction from which the shot came, leading to miscommunication and increased risk of accidental shootings. The inability to see each other and accurately locate the source of gunfire undermines the safety protocols that are crucial in hunting activities, as it impedes clear communication and coordination among the hunters in the area.

[0004] Moreover, when a target animal is shot at, it may run away from the shooter. In such scenarios, other hunters would benefit from knowing the possible direction in which the animal is fleeing. However, in a dense forest where visibility is poor and the origin of the gunshot is uncertain, it becomes almost impossible to ascertain the direction the animal is taking. This further complicates the hunting process, as hunters might miss the opportunity to track the animal or inadvertently move into unsafe areas while searching for it. The overall lack of clear auditory and visual signals in the forest creates a hazardous and inefficient hunting environment.Object of the Invention

[0005] The objective of the present invention is to provide a solution that minimizes the above-mentioned problems.Summary of the Invention

[0006] A first aspect of the present invention relates to an alert system for use during a solo hunt or a group hunt, the system comprising at least two devices, and a remote server; wherein each device comprises: a shot detection sensor component configured to sense and register if a firearm is discharged; a location sensor component configured to continuously determine the geographic location of the device; a processing component communicatively coupled to the shot detection sensor component and the location sensor, and configured to generate metadata comprising a timestamped geolocation of a registered firearm discharge, and at predefined time periods generate metadata comprising a timestamped geolocation of the device; a transmitting component communicatively coupled to the processing component and the server and configured to transmit said metadata to the server; and a receiving component communicatively coupled to the remote server; wherein the remote server is configured to: i) receive metadata from the device's transmitting component to verify the occurrence of a firearm discharge and to verify the geographic position of the device; ii) determine which users should receive alerts based on their device's current geographic locations; iii) craft a notification message that includes essential details about a verified firearm discharge event; and iv) send the crafted notification to the devices of a selection of users.

[0007] In one or more embodiments, the shot detection component comprises one or more microphones, and wherein either the shot detection component and / or the processing component is configured to employ digital signal processing techniques to preprocess audio signals registered by said microphones to extract features selected from energy, zero-crossing rate, and spectral characteristics, such as spectral centroid, bandwidth, and temporal attributes, like attack and decay times, which are indicative of firearm discharge.

[0008] In one or more embodiments, either the shot detection component and / or the processing component, based on the extracted features, is configured to determine if the registered sound is generated in proximity to the device or is distant from the device.

[0009] In one or more embodiments, either the shot detection component and / or the processing component, based on the extracted features, is configured to perform a sound classification operation to identify and categorize the audio signals captured by the device's microphone.

[0010] In one or more embodiments, the remote server is configured to, if present, aggregate metadata from a plurality of devices to verify the occurrence of a firearm discharge, and wherein the aggregation process involves collecting timestamps, GNSS coordinates, classification confidence levels, or any contextual information provided by each device.

[0011] In one or more embodiments, the aggregation process is preceded by a server check of said metadata for spatial and temporal consistency, such as comparing including comparing GNSS coordinates to confirm that the reported firearm discharges are within a close geographical range, e.g., a pre-defined hunting ground, or a pre-defined distance from a device.

[0012] In one or more embodiments, the shot detection component and / or the processing component, based on the extracted features, is configured to estimate a confidence level of a registered firearm discharge, and wherein the remote server is configured to, if present, aggregate metadata describing such confidence levels from a plurality of devices to verify the occurrence of a firearm discharge.

[0013] In one or more embodiments, the remote server is configured to continuously monitor the geolocation data of all said devices, and using this data to identify devices that are within a specific radius of a verified firearm discharge event to create a notification about said event to the devices of a selection of users based on this identification operation.

[0014] A second aspect of the present invention relates to an alert system for use during a solo hunt or a group hunt, the system comprising at least two devices, and a remote server; wherein each device comprises: a shot detection sensor component configured to sense and register if a firearm is discharged; a location sensor component configured to continuously determine the geographic location of the device; a processing component communicatively coupled to the shot detection sensor component and the location sensor, and configured to generate metadata comprising a timestamped geolocation of a registered firearm discharge, and at predefined time periods generate metadata comprising a timestamped geolocation of the device; a transmitting component communicatively coupled to the processing component and the server and configured to transmit said metadata to the server; and a receiving component communicatively coupled to the remote server; wherein the remote server is configured to: receive metadata from the device's transmitting component to verify the occurrence of a firearm discharge and to verify the geographic position of the device.

[0015] Preferably, the remote server is configured to: determine which users should receive alerts based on their device's current geographic locations.

[0016] Preferably, the remote server is configured to: craft a notification message that includes essential details about a verified firearm discharge event.

[0017] Preferably, the remote server is configured to: send the crafted notification to the devices of a selection of users.

[0018] As used in the specification and the appended claims, the singular forms "a", "an", and "the" include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from "about" or "approximately" one particular value and / or to "about" or "approximately" another particular value. When such a range is expressed, another embodiment includes from the one particular value and / or to the other particular value. Similarly, when values are expressed as approximations, by use of the antecedent "about", it will be understood that the particular value forms another embodiment.

[0019] It should be noted that embodiments and features described in the context of one of the aspects of the present invention also apply to the other aspects of the invention.Brief description of the figures

[0020] Figures 1-4 show examples of user interface views from an application running on a smartphone.Detailed Description of the Invention

[0021] In the present context, the term "in general" when used when mentioning a feature relating to the present invention, it must be understood that the feature may be used with all embodiments of the invention, even if the mentioning is made in the detailed part of the document.

[0022] In general, the device according to the present invention may be a smartphone, and the system may include a plurality of smartphones and a server. The following is a non-limiting example on how such a system and device may be constructed.User interface

[0023] In general, the user interface (Ul) of an application designed for managing the hunting activities according to the present invention would need to balance simplicity with functionality, ensuring that users can easily navigate through the different features, whether they are seasoned hunters or new to the platform.Logging in and accessing existing hunts

[0024] When a user opens the application, they may first be presented with a login screen. If they are part of an existing hunt, they can e.g., log in using credentials specific to that hunt, possibly provided by the hunt organizer. Upon successful login, the user is directed to a dashboard that displays their current hunts, upcoming hunts, and any invitations to join new hunts. An example of a user interface may be seen in Figure 1. This dashboard acts as a central hub where all relevant information is readily accessible.Creating a New Hunt

[0025] For users looking to organize their own hunt, the UI provides a straightforward process to create a new hunt (see an e.g., Figure 2). This feature might be accessed via a prominent button or menu option labelled "Create Hunt". The user may be guided through a series of steps to define the hunt's parameters, such as selecting the location, setting the date and time, specifying the types of game to be hunted, and inviting other participants. The UI may allow for customization, enabling the organizer to set rules, define zones within the hunting area, and perhaps even integrate maps to help participants navigate the terrain.Overview of Hunts

[0026] The application may also offer a comprehensive overview of all hunts, both future and historical (see e.g., Figure 3). This is typically accessible through a "Hunts" tab (see e.g., Figure 1), where hunts are categorized into upcoming and past events. Users can click on any hunt to view detailed information, such as participants, location, and rules. For past hunts, the Ul might display statistics, such as the number of animals spotted, animals shot, shots fired, and other relevant data. This feature provides users with a quick reference to their hunting history, allowing them to reflect on past experiences and plan for future hunts.Providing Feedback on Hunting Outcomes

[0027] After a hunt, the UI prompts users to provide feedback on their hunting outcomes (see e.g., Figure 4). This might be done through a post-hunt report feature, where users can log details about the animals they shot or shot at. The interface for this feature could include options to select the type of animal, the number of shots fired, whether the animal was hit or missed, and any additional notes, such as the distance of the shot or the condition of the animal. This data is then stored in the hunt's history, contributing to a collective record that can be reviewed by all participants.Overall User Experience

[0028] The Ul design should be intuitive and accessible, minimizing the number of steps required to perform each task. It should also be visually appealing, with a clean layout that emphasizes clarity and ease of use. Integrating features like real-time updates, notifications, and interactive maps can enhance the user experience, making the application a valuable tool for both individual hunters and organized hunting groups. The goal is to create an application that not only facilitates the logistics of hunting but also enhances the overall experience by providing a centralized platform for planning, execution, and reflection.On-Device Processing

[0029] The system starts with the smartphones, which are equipped to capture and analyse audio signals using their built-in microphones. Digital Signal Processing (DSP) techniques are employed to preprocess these audio signals, filtering out background noise and extracting relevant features such as energy, zero-crossing rate, and spectral characteristics. These features are critical for identifying the distinct acoustic signatures of gunshots.Sensor Data Detection

[0030] An application leverages the smartphone's microphone to detect specific audio signals corresponding to the discharge of a firearm, e.g., a specific firearm. When the microphone detects a relevant sound, the app employs an audio classifier to verify the type and intensity of the sound. This classifier can distinguish between different types of audio inputs such as speech, music, background noise, a user operating a firearm, a firearm discharge, e.g., categorizing them further by volume levels (quiet, mid-level, loud) to determine if the sound is generated in proximity to the user or is distant from the user. The app continuously captures audio using the smartphone's microphone. This may be done in a way that minimizes battery consumption, such as by periodically sampling the audio environment or by using wake-on-sound features. Such a wake-on-sound may e.g., be the user operating (e.g., grabbing or lifting) their firearm. Firearm discharge (gunshot) sounds have unique acoustic signatures characterized by a sharp onset and specific frequency patterns. The app may e.g., utilize digital signal processing (DSP) techniques to identify these characteristics. Key features may e.g., be extracted from the audio signal, such as the energy, zero-crossing rate, and spectral characteristics, which are indicative of gunshots, e.g., specific gunshots. In a smartphone-based group alert system for detecting firearm discharges and according to the present invention, both sound signature analysis and feature extraction should ideally be performed on the smartphone itself rather than on the server. Performing sound analysis and feature extraction on the device significantly reduces the latency associated with sending raw audio data to the server for processing. Real-time detection and alerts are critical in emergency situations. Audio files, especially those with high fidelity required for accurate sound analysis, can be large. Transmitting these files to a server would consume considerable bandwidth and could lead to delays. Extracting features on the device reduces the amount of data that needs to be sent. Furthermore, processing audio data on the device helps protect user privacy, as sensitive audio data does not need to be transmitted over the network. Only the extracted features or the detection metadata is sent to the server, ensuring that personal conversations or background sounds remain private. Finally, modern smartphones are equipped with powerful processors capable of handling complex computations. Utilizing these resources for on-device processing can distribute the computational load more efficiently, leaving server resources for aggregating data and managing notifications.Feature Extraction

[0031] Feature extraction in the context of detecting firearm discharges via a smartphone application involves analysing the captured audio to identify distinctive characteristics that signify a gunshot. This process starts with the smartphone's microphone continuously recording ambient sounds. The raw audio data is then subjected to digital signal processing (DSP) techniques to isolate and quantify specific features that are indicative of gunfire.

[0032] The first step in feature extraction is preprocessing the audio signal. This may involve filtering out background noise and normalizing the audio to ensure consistent analysis. The pre-processed audio may then be divided into short time frames for detailed examination, as gunshots are typically very brief events.

[0033] One key feature extracted is the energy of the sound, which measures the loudness or intensity of the audio signal. Gunshots produce high-energy bursts, making this a crucial feature. The zero-crossing rate, which counts the number of times the audio signal changes from positive to negative, is another important feature. Gunshots have a high zero-crossing rate due to their sharp and sudden onset.

[0034] Spectral features are also critical in identifying gunshots. The spectral centroid, which indicates where the centre of mass of the spectrum is located, helps in distinguishing gunshots from other sounds. Gunshots tend to have a high spectral centroid because they contain more high-frequency components. Additionally, the bandwidth and roll-off of the spectrum provide insights into the spread and decay of the frequencies in the audio signal.

[0035] Temporal features are equally important. The attack time, which is the time taken for the sound to reach its peak amplitude, and the decay time, which is the time taken for the sound to fall back to a lower amplitude, are measured. Gunshots typically have very short attack and decay times, distinguishing them from other impulsive sounds.

[0036] These extracted features may then be fed into machine learning models, such as convolutional neural networks (CNNs) or recurrent neural networks (RNNs).

[0037] These models have been trained on extensive datasets of gunshot sounds and other environmental noises, allowing them to accurately classify the audio signal.Sound classification

[0038] The sound classification in the device according to the present invention may be a sophisticated process that involves identifying and categorizing audio signals captured by the device's microphone. This process is crucial for distinguishing gunshots from other similar sounds, such as fireworks, car backfires, or construction noises.

[0039] The sound classification process may begin with the pre-processed audio data, which has undergone feature extraction to isolate relevant characteristics. As said, these characteristics, or features, may include energy, zero-crossing rate, spectral centroid, bandwidth, and temporal attributes like attack and decay times. These features may then be fed into a machine learning model specifically trained to recognize the acoustic signature of gunshots.

[0040] The machine learning model used for sound classification is typically a convolutional neural network (CNN) or a recurrent neural network (RNN). CNNs are effective because they can automatically detect spatial hierarchies in data, making them suitable for analysing the time-frequency representations of audio signals, such as spectrograms. Spectrograms visually represent the spectrum of frequencies in a sound signal as it varies with time, and CNNs can learn to identify the patterns within these spectrograms that correspond to gunshots.

[0041] RNNs, on the other hand, are well-suited for sequential data and are used when the temporal dynamics of the audio signal are crucial for classification. RNNs can consider the context provided by previous audio frames, making them effective in distinguishing gunshots from other impulsive sounds that might have similar immediate characteristics but differ in their temporal patterns.

[0042] The training process for these models involves feeding them large datasets containing various sounds, including gunshots and other environmental noises. These datasets are labelled, meaning each audio segment is annotated with the correct classification. The model learns to differentiate between these sounds by adjusting its internal parameters to minimize the error in its predictions.

[0043] Once trained, the model is deployed on the smartphone using a framework like TensorFlow Lite, which allows for efficient on-device inference. When the smartphone captures a sound, the extracted features are input to the model, which processes them and outputs a classification. If the model identifies the sound as a gunshot, it triggers the generation of metadata, including the classification confidence and geolocation, to be sent to a central server.

[0044] The central server may further validate the classification by aggregating data from multiple devices in the vicinity, ensuring that alerts are accurate and reducing the likelihood of false positives. This validation process can involve additional layers of analysis, such as cross-referencing with other detected events or corroborating reports from nearby users.

[0045] This on-device sound classification approach ensures real-time detection and reduces the need for constant data transmission, preserving bandwidth and enhancing user privacy. By leveraging advanced machine learning techniques and efficient processing frameworks, the system reliably identifies firearm discharges and promptly alerts users, significantly improving public safety.

[0046] A deep learning framework, such as TensorFlow Lite, may enable these models to run efficiently on the smartphone, ensuring real-time processing and detection. TensorFlow Lite is a lightweight, open-source deep learning framework designed for mobile and embedded devices. It is a part of TensorFlow, which is a larger machine learning library developed by Google. TensorFlow Lite enables developers to run machine learning models on resource-constrained devices such as smartphones, tablets, and embedded systems.Event metadata generation

[0047] Event metadata generation in a device according to the present invention may involve creating detailed, structured information about a detected event that can be transmitted to a central server for validation and further action. This process ensures that relevant data about the event is captured and communicated efficiently, allowing for prompt and accurate alerts to users within the affected area.

[0048] When the smartphone's sound classification system identifies a gunshot, the application immediately begins generating metadata. This metadata includes several critical pieces of information. Firstly, it records the timestamp of the event, capturing the exact time the gunshot was detected. This is essential for temporal analysis and for coordinating responses to the incident.

[0049] Next, the application retrieves the GNSS coordinates of the device at the moment of detection. These coordinates provide the precise location of the event, which is crucial for determining which users should receive alerts.

[0050] The classification confidence is another key element of the metadata. This is a numerical value that indicates how certain the machine learning model is that the detected sound is indeed a gunshot. By including the confidence level, the system can apply different thresholds for triggering alerts or deciding whether additional validation is needed. Higher confidence levels might trigger immediate alerts, while lower levels might require corroboration from other nearby devices.

[0051] In addition to these core elements, the metadata may also include contextual information about the environment at the time of detection. This can involve data from other sensors on the smartphone / device, such as accelerometers, which might indicate whether the phone was stationary or in motion, or ambient light sensors, which could provide information about the lighting conditions. Such contextual data helps in understanding the circumstances of the event and can be useful for further analysis and decision-making.

[0052] Once the metadata is compiled, it is securely transmitted to the central server. The use of metadata, rather than raw audio data, minimizes bandwidth usage and enhances user privacy by ensuring that personal and sensitive audio content is not exposed.Server-side aggregation and validation

[0053] Server-side aggregation and validation are critical components in a system according to the present invention, ensuring that the alerts sent to users are accurate and reliable. This process begins once the server receives metadata from multiple devices / smartphones that have detected potential gunshot events.

[0054] When the central server receives the metadata, it aggregates data from various sources to verify the occurrence of a firearm discharge. The aggregation involves collecting timestamps, GNSS coordinates, classification confidence levels, and any contextual information provided by each smartphone. By pooling this information, the server can create a comprehensive view of the event.

[0055] The server first checks for spatial and temporal consistency among the reports. It compares the GNSS coordinates to confirm that the reported gunshots are within a close geographical range, e.g., a pre-defined hunting ground, or a pre-defined distance from a user. If multiple devices report a gunshot from the same area around the same time, this increases the likelihood that a real event has occurred. The server may use clustering algorithms to group reports that are spatially and temporally proximate.

[0056] Next, the server evaluates the classification confidence levels provided by each smartphone. If several devices report high confidence levels, the server can be more certain of the gunshot's authenticity. Conversely, if confidence levels are low or vary significantly, the server might flag the event, e.g., for further review, or seek additional corroboration before issuing an alert.

[0057] The server may also consider contextual information, such as movement detected by accelerometers, ambient light conditions, and any other relevant data included in the metadata. This additional information can help in distinguishing true gunshot events from false positives caused by other loud noises or environmental factors.

[0058] To further enhance validation, the server may cross-reference the aggregated data with external sources, such as other surveillance systems in the area. This multi-source verification helps confirm the gunshot's occurrence and provides additional context for the event.Group notification

[0059] Group notification is a crucial aspect of a system according to the present invention, ensuring that timely and accurate alerts are delivered to users in the vicinity of the detected event. This process leverages real-time geolocation data and sophisticated notification mechanisms to enhance public safety.

[0060] Once the server has validated the occurrence of a firearm discharge through aggregation and analysis of metadata from multiple smartphones, it determines which users should receive alerts based on their current locations. The server may continuously monitor the geolocation data of all users who have the application installed. Using this data, the server identifies users who are within a specific radius of the confirmed gunshot event. This geographical filtering ensures that only those who are interested in the incident are notified.

[0061] The server then crafts a notification message that includes essential details about the gunshot event. This message may typically contain the location of the incident, the time it was detected, and any recommended actions. The content of the notification is designed to be concise yet informative, providing users with the necessary information to make immediate safety decisions, or to know who shot, or shot at, a target animal.

[0062] To deliver these notifications, the system utilizes push notification services, such as Firebase Cloud Messaging (FCM) for Android devices and Apple Push Notification Service (APNs) for iOS devices. These services enable the server to send real-time alerts directly to users' smartphones. The notifications appear on the users' screens, even if the application is not actively running, ensuring that the alert is seen promptly.

[0063] The notification process is optimized to handle high volumes of alerts efficiently, maintaining performance and reliability even during large-scale incidents. The system can also prioritize notifications based on the events closest to a user.

[0064] To further enhance the effectiveness of group notifications, the system can incorporate user preferences and feedback. Users can customize their notification settings, choosing to receive alerts for specific types of events or adjusting the sensitivity of the notifications, e.g., based on their personal risk tolerance. Additionally, the system can use feedback from users to refine the notification process, learning from past incidents to improve accuracy and responsiveness.

[0065] Group notification in this context is not just about delivering messages but also about fostering a hunter / user response to emergencies. By ensuring that accurate and timely information is disseminated to those who need it most, the system empowers users to take proactive measures to protect themselves and assist others.Examples on implementation details

[0066] Backend Infrastructure: Implement using scalable cloud services that can handle real-time data processing and push notifications. Technologies like AWS Lambda, Firebase, or a microservices architecture on Kubernetes can be employed. Real-Time Communication: Use WebSockets or push notification services (e.g., Firebase Cloud Messaging) for immediate data relay. Data Processing: Employ e.g., machine learning models for audio classification, which can be trained on various audio datasets. TensorFlow Lite can be integrated into the app for on-device inference to reduce latency. Security: Ensure all data transmissions are encrypted using protocols such as TLS. User consent and privacy policies should be strictly adhered to, with options for users to control their data sharing preferences. By integrating one or more of these components, the application provides a robust platform for real-time, context-aware communication, enhancing safety, and collaboration among users based on their geographical and environmental context.

[0067] WebSockets is a communication protocol that provides full-duplex communication channels over a single, long-lived TCP connection. It is designed to facilitate real-time communication between a client (such as a web browser) and a server. Unlike the traditional request-response model of HTTP, WebSockets allow for a persistent connection where both the client and server can send messages to each other at any time.

Claims

1. An alert system for use during a solo hunt or a group hunt, the system comprising at least two devices, and a remote server; wherein each device comprises: - a shot detection sensor component configured to sense and register if a firearm is discharged; - a location sensor component configured to continuously determine the geographic location of the device; - a processing component communicatively coupled to the shot detection sensor component and the location sensor, and configured to generate metadata comprising a timestamped geolocation of a registered firearm discharge, and at predefined time periods generate metadata comprising a timestamped geolocation of the device; - a transmitting component communicatively coupled to the processing component and the server and configured to transmit said metadata to the server; and - a receiving component communicatively coupled to the remote server; wherein the remote server is configured to: i) receive metadata from the device's transmitting component to verify the occurrence of a firearm discharge and to verify the geographic position of the device; ii) determine which users should receive alerts based on their device's current geographic locations; iii) craft a notification message that includes essential details about a verified firearm discharge event; and iv) send the crafted notification to the devices of a selection of users.

2. The alert system according to claim 1, wherein the shot detection component comprises one or more microphones, and wherein either the shot detection component and / or the processing component is configured to employ digital signal processing techniques to preprocess audio signals registered by said microphones to extract features selected from energy, zero-crossing rate, and spectral characteristics, such as spectral centroid, bandwidth, and temporal attributes, like attack and decay times, which are indicative of firearm discharge.

3. The alert system according to claim 2, wherein either the shot detection component and / or the processing component, based on the extracted features, is configured to determine if the registered sound is generated in proximity to the device or is distant from the device.

4. The alert system according to any one of the claims 2-3, wherein either the shot detection component and / or the processing component, based on the extracted features, is configured to perform a sound classification operation to identify and categorize the audio signals captured by the device's microphone.

5. The alert system according to any one of the claims 1-4, wherein the remote server is configured to, if present, aggregate metadata from a plurality of devices to verify the occurrence of a firearm discharge, and wherein the aggregation process involves collecting timestamps, GNSS coordinates, classification confidence levels, or any contextual information provided by each device.

6. The alert system according to claim 5, wherein the aggregation process is preceded by a server check of said metadata for spatial and temporal consistency, such as comparing including comparing GNSS coordinates to confirm that the reported firearm discharges are within a close geographical range, e.g., a pre-defined hunting ground, or a pre-defined distance from a device.

7. The alert system according to any one of the claims 2-6, wherein the shot detection component and / or the processing component, based on the extracted features, is configured to estimate a confidence level of a registered firearm discharge, and wherein the remote server is configured to, if present, aggregate metadata describing such confidence levels from a plurality of devices to verify the occurrence of a firearm discharge.

8. The alert system according to any one of the claims 1-7, wherein the remote server is configured to continuously monitor the geolocation data of all said devices, and using this data to identify devices that are within a specific radius of a verified firearm discharge event to create a notification about said event to the devices of a selection of users based on this identification operation.

Images