Watercraft having a perception-based sensor

Abstract

A watercraft equipped with at least one perception-based sensor located behind the front splash guard, in front of the steering handlebar, and above a waterline of the watercraft is disclosed.

Description

[0001] WATERCRAFT HAVING A PERCEPTION-BASED SENSOR

[0002] CROSS-REFERENCE SECTION

[0003] [1] This international application claims priority to US Provisional Patent Application No. 63 / 701,457, filed 09 / 30 / 2024, titled "WATERCRAFT HAVING A PERCEPTION-BASED SENSOR TECHNICAL FIELD”, the entire contents of which are incorporated herein by reference.

[0004] TECHNICAL FIELD

[0005] [2] The present technology relates to the technical field of watercraft, specifically to watercraft having perception-based sensor.

[0006] BACKGROUND

[0007] [3] In the field of trajectory determination systems, traditional approaches employ optical sensors to detect nearby vehicles and objects. These systems are commonly utilized in autonomous vehicles navigating on relatively flat level road surfaces. The pitch and roll angles of such vehicles have minimal impact on trajectory calculations due to their stability.

[0008] [4] The automotive industry has adopted technologies like C-V2X to facilitate communication between vehicles and their surroundings. This technology enables vehicles to exchange information with other devices in real-time, enhancing safety and efficiency. Nevertheless, a major limitation of these technologies is that both communicating parties must be equipped with C-V2X for interaction. As a result, a newer vehicle with C-V2X cannot communicate with an older vehicle lacking this technology.

[0009] [5] Despite the advantages of C-V2X in enhancing road safety and efficiency, its limitation of requiring both parties to have the technology on-board poses a significant challenge. In scenarios where older vehicles lack C-V2X capabilities, communication is hindered.

[0010] 303774181.1 [6] These solutions used in the automotive industry are difficult to implement on a watercraft often pitch, roll and yaw on water, all of which can affect the trajectory of the watercraft.

[0011] [7] It is therefore an objective of the present technology to overcome at least in part the aforementioned limitations.

[0012] SUMMARY

[0013] [8] The present technology has been designed to overcome at least some drawbacks present in prior art solutions.

[0014] [9] According to an aspect, the present technology relates to a watercraft comprising: a) a hull; b) a deck mounted on the hull; c) a motor connected to at least one of the deck and the hull; d) a propulsion system operatively connected to the motor; e) a steering handlebar operatively connected to the deck; f) a front splash guard mounted to the deck forward of the steering handlebar; g) a perception-based sensor mounted: i) behind the front splash guard; ii) in front of or on the steering handlebar; and iii) above a waterline of the watercraft

[0015] 303774181.1

[0010] Before providing below a detailed review of embodiments of the technology, some optional characteristics that may be used in association or alternatively will be listed hereinafter:

[0016]

[0011] According to an example, the perception-based sensor is located on a front cowling of the watercraft.

[0017]

[0012] According to an example, the present technology further comprises a housing housing the perception-based sensor, the housing being configured to protect the perception-based sensor from water.

[0018]

[0013] According to an example, the perception-based sensor is an optical sensor.

[0019]

[0014] According to an example, the perception-based sensor is a camera capable of capturing at least one of images or videos in real time.

[0020]

[0015] According to an example, the present technology further comprises a processing module connected to the perception-based sensor, the processing module being responsible for analyzing data from the perception-based sensor and for determining the presence and location of objects in a vicinity of the watercraft.

[0021]

[0016] According to an example, the present technology comprises a system for perceptionbased trajectory determination connected to the perception-based sensor.

[0022]

[0017] According to an example, the system for perception-based trajectory determination for a watercraft comprises: a) a neural network module comprising a pre-trained neural network and being configured to detect an object within a predetermined area, the predetermined area being located around the watercraft, using the acquired data; b) a processing module configured to: i) acquire data from the perception-based sensor; ii) send the acquired data to the neural network module;

[0023] 303774181.1 iii) detect an object located in the predetermined area of observation using acquired data from the perception-based sensor; iv) extract a first set of information regarding the detected object, the first set of information being configured to be used to determine a predicted trajectory of the detected object, the first set of information comprising:

[0024] (1) a relative distance between the detected object and the watercraft;

[0025] (2) a detected object position based on the coordinates of the detected object within a field of view of the perception-based sensor; and

[0026] (3) a relative bearing between the detected object and the watercraft; v) acquire a second set of information about the watercraft from a set of sensors of the watercraft; vi) compute a third set of information using the first set of information and the second set of information secondary information, the third set of information comprising a predicted trajectory of the detected object and a predicted trajectory of the watercraft and a relative speed between the detected object and the watercraft; vi i J determine a closest distance between the detected object and the watercraft exist, based on the predicted trajectory of the detected object, the predictive trajectory of the watercraft and the relative speed between the detected object and the watercraft; and viii) in response to the determination of the closest distance, execute a safety action among a set of safety actions according to a predetermined set of safety rules.

[0027]

[0018] According to an example, the processing module is further configured to provide a user of the watercraft with real-time information about the detected object.

[0028]

[0019] According to an example, the safety action is taken from a set of safety actions according to a predetermined set of safety rules.

[0029] 303774181.1

[0020] According to an example, in response to the determination of the closest distance, the processing module is configured to determine at least one of: a) a direct distance between the actual position of the watercraft and a point of the predicted trajectory of the watercraft corresponding to the point of the closest distance; or b) a time to the closest distance corresponding to the time to reach a point of the predicted trajectory of the watercraft corresponding to the point of the closest distance.

[0030]

[0021] According to an example, the processing module is further configured to: a) warn the user via a notification in response to at least one of the time to the closest distance being smaller than a predetermined first time threshold or the direct distance being smaller than a first predetermined distance threshold; b) reduce motor power output below a predetermined limit in response to at least one of the time to the closest distance being smaller than a predetermined second time threshold or the direct distance being smaller than a second predetermined threshold; and c) automatically brake the watercraft using a reverse gate in response to at least one of the time to the closest distance being smaller than a predetermined third time threshold or the direct distance being smaller than a third predetermined threshold.

[0031]

[0022] According to an example, the perception-based sensor comprises an additional camera, and the camera comprises a first field of view, and the additional camera comprises a second field of view, the first field of view being wider than the second field of view.

[0032]

[0023] According to an example, the perception-based sensor is mounted to form an angle Al with the waterline, the angle Al being comprised between -25 degrees and +25 degrees.

[0033]

[0024] According to an example, the first set of information comprises an estimation of the detected object attitude comprising an estimation of at least one of: a) a pitch of the detected object;

[0034] 303774181.1 b) a roll of the detected object; or c) a yaw of the detected object.

[0035]

[0025] According to an example, the present technology comprises a water jet propulsion system for generating a thrust, and a reverse gate configured to redirect the thrust toward a front of the watercraft.

[0036] BRIEF DESCRIPTION OF THE DRAWINGS

[0037]

[0026] For a better understanding of the present technology, as well as other aspects and further features thereof, reference is made to the following description which is to be used in conjunction with the accompanying drawings, where:

[0038]

[0027] Figure 1 is a right-side elevation view of a personal watercraft;

[0039]

[0028] Figure 2 is a top-plan view of the watercraft of Figure 1;

[0040]

[0029] Figure 3 is a flowchart representing steps involved in determining perception trajectories;

[0041]

[0030] Figure 4 is a schematic representation of a system for determining perception trajectories;

[0042]

[0031] Figure 5 is a right-side elevation view of the watercraft of Figure 1 equipped with a perception-based sensor;

[0043]

[0032] Figure 6 illustrates the correlation between detected objects and their computed positions;

[0044]

[0033] Figure 7 illustrates a calculated distance between a detected object and a watercraft;

[0045]

[0034] Figure 8 is a perspective view taken from a front, right side of a personal watercraft showing a placement zone for a perception-based;

[0046]

[0035] Figure 9 illustrates a schematic layout of a perception-based sensor's field of view on the watercraft of Figure 8;

[0047] 303774181.1

[0036] Figure 10 illustrates a schematic layout of a dual-field-of-view configuration of a perception-based sensor on the watercraft of Figure 8;

[0048]

[0037] Figure 11 illustrates a secured region in front of a watercraft ;

[0049]

[0038] Figure 12 schematically illustrates a docking process of a watercraft;

[0050]

[0039] Figure 13 illustrates a triangular arrangement of observational elements of a watercraft;

[0051]

[0040] Figure 14 is a flow diagram of a method for calculating a throttle limitation;

[0052]

[0041] Figure 15 illustrates an implementation process of the method of Figure 14;

[0053]

[0042] Figure 16 illustrates a calculation process of throttle limitation;

[0054]

[0043] Figure 17 illustrates predicted trajectories of two watercraft;

[0055]

[0044] Figure 18 illustrates a flow diagram of a sub-method configured to determine at least one limitation parameter using tables; and

[0056]

[0045] Figure 19 illustrates a flow diagram of a sub-method to filter some parameters.

[0057] DETAILED DESCRIPTION

[0058]

[0046] The examples and conditional language recited herein are principally intended to aid the reader in understanding the principles of the present technology and not to limit its scope to such specifically recited examples and conditions. It will be appreciated that those skilled in the art may devise various arrangements which, although not explicitly described or shown herein, nonetheless embody the principles of the present technology and are comprised within its spirit and scope.

[0059]

[0047] Furthermore, as an aid to understanding, the following description may describe relatively simplified implementations of the present technology. As persons skilled in the art would understand, various implementations of the present technology may be of a greater complexity.

[0060] 303774181.1

[0048] In some cases, what are believed to be helpful examples of modifications to the present technology may also be set forth. This is done merely as an aid to understanding, and, again, not to define the scope or set forth the bounds of the present technology. These modifications are not an exhaustive list, and a person skilled in the art may make other modifications while nonetheless remaining within the scope of the present technology. Further, where no examples of modifications have been set forth, it should not be interpreted that no modifications are possible and / or that what is described is the sole manner of implementing that element of the present technology.

[0061]

[0049] Moreover, all statements herein reciting principles, aspects, and implementations of the present technology, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof, whether they are currently known or developed in the future. Thus, for example, it will be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative circuitry embodying the principles of the present technology. Similarly, it will be appreciated that any flowcharts, flow diagrams, state transition diagrams, pseudo-code, and the like represent various processes which may be substantially represented in computer-readable media and so executed by a computer or processor, whether or not such computer or processor is explicitly shown.

[0062]

[0050] Unless otherwise specified herein, or unless the context clearly dictates otherwise the term about modifying a numerical quantity means plus or minus ten percent. Unless otherwise specified, or unless the context dictates otherwise, between two numerical values is to be read as between and including the two numerical values.

[0063]

[0051] In the present description, some specific details are comprised to provide an understanding of various disclosed implementations. The skilled person in the relevant art, however, will recognize that implementations may be practiced without one or more of these specific details, parts of a method, components, materials, etc. In some instances, well-known methods associated with artificial intelligence, machine learning and / or neural networks, have not been shown or described in detail to avoid unnecessarily obscuring descriptions of the disclosed implementations.

[0064]

[0052] In the present description and appended claims "a", "an", "one", or "another" applied to "embodiment", "example", or "implementation" is used in the sense that a particular

[0065] 303774181.1 referent feature, structure, or characteristic described in connection with the embodiment, example, or implementation is comprised in at least one embodiment, example, or implementation. Thus, phrases like "in one embodiment", "in an embodiment", or "another embodiment" are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics maybe combined in any suitable manner in one or more embodiments, examples, or implementations.

[0066]

[0053] As used in this description and the appended claims, the singular forms of articles, such as "a", "an", and "the", may comprise plural referents unless the context mandates otherwise. Unless the context requires otherwise, throughout this description and appended claims, the word "comprise" and variations thereof, such as, "comprises" and "comprising" are to be interpreted in an open, inclusive sense, that is, as "including, but not limited to".

[0067]

[0054] In the context of the present technology, "neural network”, equally referred to "trained neural network” or "pre-trained neural network” may refer to a type of artificial intelligence that uses a computer system to perform calculations, simulations, or predictions based on mathematical formulas or rules. A neural network may be used to solve complex problems that require high-level abstraction, reasoning, or computation. A neural network may be created from scratch or based on an existing model that has been trained on a dataset. A neural network may have various applications, such as trajectory determination, natural language processing, computer vision, information retrieval, or data analysis.

[0068]

[0055] In the context of the present technology, "data” may refer to a representation of information that may be processed, stored, transmitted, or manipulated by a computer system. A data may be a document, a picture, a video, a multimedia file, a text, information related to an object, a device or to a product, etc. A data may have different formats, such as binary, text, audio, image, or video. A data may also have different attributes, such as size, type, content, or metadata. A data may be used for various purposes, such as communication, analysis, or computation, for example.

[0069]

[0056] In the context of the present specification, a "database” is any structured collection of data, irrespective of its particular structure, the database management software, or the computer hardware on which the data is stored, implemented or otherwise rendered available for use. A database may reside on the same hardware as the process that stores or makes use of the information stored in the database or it may reside on separate hardware,

[0070] 303774181.1 such as a dedicated server or plurality of servers. It may be that a database is a logically ordered collection of structured data kept electronically in a computer system.

[0071]

[0057] In the context of the present specification, the expression "information” includes information of any nature or kind whatsoever capable of being stored in a database.

[0072]

[0058] The functions of the various elements shown in the figures, including any functional block labeled as a "processor", "processing unit”, "processing module” or a "graphics processing unit”, may be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functions may be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be shared. In some embodiments of the present technology, the processor may be a general purpose processor, such as a central processing unit (CPU) or a processor dedicated to a specific purpose, such as a graphics processing unit (GPU). Moreover, explicit use of the term "processor" or "controller" should not be construed to refer exclusively to hardware capable of executing software, and may implicitly include, without limitation, digital signal processor (DSP) hardware, network processor, application specific integrated circuit (ASIC), field programmable gate array (FPGA), read-only memory (ROM) for storing software, random access memory (RAM), and non-volatile storage. Other hardware, conventional and / or custom, may also be included.

[0073]

[0059] Software modules, or simply modules which are implied to be software, may be represented herein as any combination of flowchart elements or other elements indicating performance of process steps and / or textual description. Such modules may be executed by hardware that is expressly or implicitly shown.

[0074]

[0060] In the context of the present specification, the expression "computer usable information storage medium” or "non-volatile memory” are intended to include media of any nature and type whatsoever, including RAM, ROM, disks (CD-ROMs, DVDs, floppy disks, hard drivers, etc.), USB keys, solid state-drives, tape drives, etc.

[0075]

[0061] With these fundamentals in place, we will now consider some non-limiting examples to illustrate various implementations of aspects of the present technology.

[0076] 303774181.1

[0062] According to an embodiment of the present technology, a system and method perception-based trajectory determination for watercraft is provided. This system utilizes a perception-based sensor on the watercraft to detect other watercraft, objects, or obstacles in its vicinity. The data from this sensor is analyzed using a neural network to determine the relative distance between the detected object and the watercraft, the object position based on the coordinates of the object within a field of view of the perception-based sensor, and the relative bearing between the detected object and the watercraft. According to an example, the data from this sensor can also be analyzed using a neural network to determine the distance, attitude, relative velocity, and relative bearing between the watercraft and the detected object. According to another embodiment, classical methods can be used to determine the relative distance between the detected object and the watercraft, the object position based on the coordinates of the object within a field of view of the perception-based sensor, and the relative bearing between the detected object and the watercraft. According to another example, classical methods can also be used to determine the distance, attitude, relative velocity, and relative bearing between the watercraft and the detected object. According to an embodiment, classical methods can comprise sophisticated approaches like stereo vision, monocular cues, feature-based detection, and / or edge / shape-based methods, for example through various mathematical transformations. These sophisticated approaches can be used for better accuracy and understanding in computer vision tasks.

[0077]

[0063] Indeed, the use of watercraft for recreational activities, for example, is becoming more popular and with increased objects on the water, being able to detect and avoid these objects becomes a useful feature for users looking for increased safety with this kind of activity.

[0078]

[0064] As represented in figures 1, 2 and 5, a watercraft 10 has a hull 80 and a deck 81 mounted to the hull 80. A straddle seat 85 is connected to the deck 81. A motor 82 is connected to the hull 80. A jet propulsion system 83 is connected to the hull 80 and is driven by the motor 82. A reverse gate 86 is connected to the rear of the watercraft 10 near the jet propulsion system 83. A handlebar 84 is pivotally mounted to the deck 81 forward of the straddle seat 85. A front splashguard 50 is mounted to the deck 81 forward of the handlebar 84. A front cowling 87 is connected to the deck 81 longitudinally between the handlebar 84 and the splashguard 50. A screen 231 is provided forward ofthe handlebar 84 and is partially

[0079] 303774181.1 housed by the front cowling 87. The screen 231 provides information to the driver such as vehicle speed, as well as other information as will be described in greater detail below.

[0080]

[0065] The hull 80 is the main body of the watercraft 10 that provides buoyancy, allowing it to float on water. It is typically designed to be streamlined to reduce resistance and improve stability and speed. The hull 80 forms the foundation of the watercraft 10 and can vary in shape and size depending on the type and purpose of the watercraft 10.

[0081]

[0066] The deck 81 is mounted on top of the hull 80, providing a platform for passengers, cargo, and equipment. It serves as the operational area of the watercraft 10 where various controls and functional elements are located.

[0082]

[0067] The watercraft 10 is powered by a motor 82, which provides the necessary power to propel the watercraft 10 through the water. The motor 82 can be an internal combustion engine or an electric motor, and it is connected to the propulsion system 83 to drive the watercraft 10 forward. In the present embodiment, the motor 82 is connected to the hull 80, but is it contemplated that it could be connected to the deck 81.

[0083]

[0068] The jet propulsion system 83 is responsible for propelling the watercraft 10 through the water. The jet propulsion system 83 has a jet pump, a venturi, and a steering nozzle (all not shown). It is contemplated that jet propulsion system 83 could be replaced by another type of propulsion system depending on the type of watercraft and its intended use. For example, some watercraft could be propelled by an outboard motor having a propeller assembly.

[0084]

[0069] The handlebar 84 allows the driver, i.e., the user, to control the direction of the watercraft 10. In the present embodiment, turning the handlebar 84 turns the steering nozzle of the jet propulsion system 83, thereby redirecting the thrust generated by the jet propulsion system 83. It is contemplated that in some embodiment, the handlebar 84 could be replaced by a steering wheel, a joystick or some other steering input device. It is also contemplated that depending on the type of propulsion system, steering could be achieved differently. For example, steering could be achieved by turning a rudder or by pivoting a propeller.

[0085]

[0070] The front splash guard 50 is a protective barrier mounted at the front of a watercraft 10. Its primary function is to shield the operator and passengers from water spray that may

[0086] 303774181.1 be kicked up while navigating through water. By deflecting water away from the occupants, the splash guard helps maintain visibility and comfort.

[0087]

[0071] A watercraft 10 may include various additional features depending on its design and purpose. These can include storage compartments, safety equipment, navigation aids, and protective elements such as windshields.

[0088]

[0072] The reverse gate 86 selectively redirects the flow of water expelled from the jet propulsion system 83 toward the front of the watercraft 10. The reverse gate 86 can move between various positions, including a stowed position, a neutral position, and a fully lowered position. The stowed position is where the reverse gate 86 does not interfere with the jet of water expelled from the propulsion system 83. The neutral position redirects the water downward, providing no significant forward or backward thrust, and the fully lowered position redirects the water forward, creating reverse thrust This reverse thrust can be used to make the watercraft 10 move backward, and in certain specifically designed systems, to decelerate / brake the watercraft 10. It is contemplated that in some embodiments, the reverse gate 86 could be omitted. Furthermore, the reverse gate 86 can be moved in between these various positions. For example, to perform "Thrust Vectoring” for a force based control system of the vehicle.

[0089]

[0073] For example, in response to a deceleration signal, a processing module sends a command to lower the reverse gate 86 towards the deceleration position. Initially, the motor speed can be reduced to avoid damage or excessive force on the reverse gate 86 as it moves.

[0090]

[0074] As the reverse gate 86 moves from the stowed position towards the fully lowered position, it may pass through an intermediate position. In this intermediate state, the thrust request may be increased or lowered slightly to prepare for the full deceleration thrust needed once the reverse gate is fully lowered.

[0091]

[0075] In response to the reverse gate 86 reaching the fully lowered position, the waterjet is redirected forward, creating a reverse thrust. This action slows down the watercraft 10 significantly, and if the deceleration device continues to be actuated, the watercraft 10 can even move in reverse.

[0092]

[0076] This allows for precise control over the watercraft's speed and direction, utilizing the reverse gate 86 for deceleration and maneuverability in water environments.

[0093] 303774181.1

[0077] According to an embodiment of the present technology, the watercraft 10 comprises aperception-basedsensor 210. As illustrated by figures 1, 5, 8, 9 and 10, the perception-based sensor 210 is mounted on the watercraft's front cowling 87, with an area designed to receive the perception-based sensor 210 being located above a waterline 60 of the watercraft 10. A perception-based sensor 210 is a type of sensor designed to gather information about the surrounding environment by interpreting various physical phenomena. These sensors operate based on different modalities such as optical signals or radio waves or light pulses. Examples include optical sensors such as cameras, which capture visual information in the form of images or videos; laser imaging detection and ranging (LiDAR) sensors, which use laser pulses to measure distances and create detailed 3D maps; and radar sensors, which emit radio waves to detect objects and their velocity.

[0094]

[0078] According to an embodiment, the perception-based sensor 210 is mounted behind the front splash guard 50. Indeed, the perception-based sensor is positioned behind the front splash guard 50 of the watercraft 10 to shield it from water and debris. This placement not only protects the sensor 210 but also allows it to have an unobstructed view of the environment in front of the watercraft 10.

[0095]

[0079] According to an embodiment, the perception-based sensor 210 is mounted in front of or on the handlebar 84, providing a clear field of view of the environment in front of the watercraft 10. Indeed, when the perception-based sensor is positioned on the handlebar it allows its Field-Of-View(FOV) to be aligned with the driver directional intent.

[0096]

[0080] According to an embodiment, the perception-based sensor 210 can be mounted along a longitudinal centerline 70 of the watercraft 10. This placement ensures that the sensor's detection capabilities are accurate and reliable since it is directly aligned with the watercraft's path, for example. Equivalently the perception-based sensor can be mounted in an off-centered position, such as in one of the rearview mirrors.

[0097]

[0081] According to an embodiment, the perception-based sensor 210 is mounted above a waterline 60 of the watercraft 10. Mounting the perception-based sensor 210 above the waterline 60 helps to prevent sensor damage, but more importantly water droplets forming on or in front of the sensor.

[0098] 303774181.1

[0082] According to an embodiment, the perception-based sensor 210 is mounted on the front cowling 87 of the watercraft 10, allowing for direct line-of-sight observation of the water surface. For example, the perception-based sensor can be mounted at least 20 inches above the waterline.

[0099]

[0083] According to an embodiment, and as represented in figure 5, the perception-based sensor 210 is mounted to form an angle Al with the waterline 60, Al being comprised between -25 degrees and +25 degrees

[0100]

[0084] According to an embodiment, the perception-based sensor 210 can be mounted on the steering wheel. This can, for example, allow the utilization of a narrower FO V sensor that follows driver directional intent.

[0101]

[0085] For example, the FOV of the perception-based sensor 210 can be around 120 degrees.

[0102]

[0086] According to an embodiment, the perception-based sensor 210 is located higher than the bumper and lower than the top of the handlebar 84.

[0103]

[0087] According to an embodiment, the watercraft 10 comprises a housing for the perception-based sensor 210. This housing is constructed to shield the perception-based sensor 210 from the surrounding water. The housing can be watertight and sealed to prevent any ingress of water. The material used for making the housing can be resistant to water and corrosion and is anti-sparking. The housing design allows easy access to the perceptionbased sensor 210 for maintenance or replacement purposes.

[0104]

[0088] The housing may be attached to the watercraft 10 using suitable mounting means, such as screws or bolts or adhesive or directly molded into the cowling. Furthermore, the housing may be transparent or have a transparent window to enable visual inspection of the perception-based sensor's surroundings.

[0105]

[0089] The size and shape of the housing depend on the specific requirements of the application and the dimensions of the perception-based sensor 210. The housing may be integrated into the watercraft's design or attached as an external component.

[0106]

[0090] According to an embodiment, and as illustrated by figure 9, the perception-based sensor 210 comprises a camera, the camera comprising a first field of view 211. According to

[0107] 303774181.1 another embodiment, and as illustrated by figure 10, the perception-based sensor 210 comprises the camera and an additional camera, the additional camera comprising a second field of view 212 narrower than the first field of view. According to an embodiment, the perception-based sensor 210 can comprise a stereo camera. According to an embodiment, the primary function of the camera can be to capture a broad perspective, while the additional camera, with its narrower field of view, can serve a complementary role. This design allows for more comprehensive coverage and potential for improved image processing or analysis. For example, a large field of view is useful to detect close objects and for fast safety response. For example, a narrow field of view allows to concentrate pixels and to get a better resolution for objects far away, this allows to anticipate safety response at long range. According to an embodiment, the field of view of the perception-based sensor can comprise a cone of perception including angle ("width") and radius ("depth") and being dynamic. It can be computed and adjusted based on riding modes (e.g. sport), current watercraft speed, current trajectory, heading, etc., for example. Same is also true for a cone orientation i.e. angle formed by center of cone and watercraft centerline. Generally, the cone orientation is coincident with centerline, for example. According to an embodiment, multiples cones can be used to cover different fields and depths of views when using multiple sensors for example.

[0108]

[0091] For example, the wider angle of the first field of view enables it to capture larger scenes or areas, while the additional camera's narrower field of view focuses on specific details within the broader scene. This dual-camera setup can lead to enhanced measurements.

[0109]

[0092] According to an embodiment, the perception-based sensor 210 is configured to collect at least one of images or videos. These images may represent different aspects of a subject or scene and can be in various formats, including, but not limited to, JPEG, PNG, or TIFF.

[0110]

[0093] According to an embodiment, processing the image data allows for the extraction of valuable information that can be used for trajectory predictions, as described hereafter. For instance, image recognition algorithms can identify objects or patterns within the images, while image compression techniques can reduce the size of the data for efficient storage and transmission.

[0111] 303774181.1

[0094] Moreover, image processing techniques such as filtering, segmentation, and enhancement can improve the quality of the images, making them more suitable for further analysis.

[0112]

[0095] Additionally, machine learning models trained on large datasets of labeled images can be employed to classify or recognize specific features within new images. Moreover, the system may employ methods of image analysis that focus on higher priority image zones to be used to classify or recognize the above features.

[0113]

[0096] According to an embodiment, the perception-based sensor 210 can also comprise a LiDar or a radar. Employing LiDar or radar as the perception-based sensor 210 offers several technical advantages. LiDar systems can measure distances by illuminating the target with laser light and measuring the time it takes for the reflected light to return. This method is highly accurate and can be used to create detailed 3D maps of environments. Radar, on the other hand, uses radio waves to detect objects and their distance. It operates based on the principle of sending out electromagnetic waves and measuring the time taken for the waves to bounce back from an object. Radar systems are particularly useful in environments with low visibility or where LiDar may not function correctly due to heavy rain, fog, or bright sunlight. Additionally, radars can detect objects that are not directly in the sensor's line of sight and can provide information on their velocity and direction. Incorporating a LiDar or radar into the present technology allows for versatile and robust object detection and distance measurement capabilities. The choice between LiDar and radar and camera depends on the specific application requirements, such as environmental conditions, desired accuracy, and operational range.

[0114]

[0097] According to an embodiment, and as represented by figure 4, the watercraft 10 comprises a system 200 for perception-based trajectory determination. This system 200 comprises the perception-based sensor 210.

[0115]

[0098] According to an embodiment, and as described in more detail hereafter, the system 200 comprises a neural network module 220. This neural network module 220 comprises a pre-trained neural network. This pre-trained neural network is configured to detect an object 20 within a predetermined area, the predetermined area being located in front of the watercraft 10, using the acquired data. In figure 11, a portion 13 of this predetermined area of observation is represented. According to an embodiment, the neural network module 220

[0116] 303774181.1 can be part of the processing module 230. According to an embodiment, the neural network module can comprise at least one processing unit and / or at least one graphical processing unit.

[0117]

[0099] According to an embodiment, the system 200 comprises a processing module 230 configured to execute a plurality of actions described hereafter. The processing can comprise at least one processing unit and / or at least one graphical processing unit. The processing module 230 can be integrated with other systems on the watercraft 10 to provide real-time information about the surrounding environment. For instance, it can be used to help collision avoidance or navigation systems. The processing module 230 may employ various algorithms and machine learning techniques to improve object detection accuracy and reliability, as described hereafter.

[0118]

[0100] Additionally, the system 200 may comprise one or more processors, memory units, and communication interfaces. These components enable the system 200 to process data, store information, and transmit data to other devices or systems as needed.

[0119]

[0101] Moreover, the system 200 can be designed to operate in standalone mode or as part of a larger network. This flexibility allows it to function independently or collaborate with other systems to provide more comprehensive solutions.

[0120]

[0102] For example, the system 200 can be powered by an internal battery or an external power source.

[0121]

[0103] The system 200 can be implemented using known technologies and components, such as micro-processing modules, actuators, sensors, and communication interfaces. The design and functionality of the system 200 can be optimized to reduce power consumption and enhance reliability and durability.

[0122]

[0104] According to an embodiment, the watercraft 10 can also comprise a set of sensors. According to an embodiment, the set of sensors comprises a Global Positioning System (GPS) 240 and / or an Inertial Measurement Unit (IMU) 250. The GPS 240 provides location information based on satellite signals. The IMU 250, on the other hand, measures and reports angular velocity and acceleration to determine vehicle attitude, acceleration, and motion. Incorporating both a GPS 240 and an IMU 250 in the set of sensors enhances positioning accuracy and improves system 200 reliability. Furthermore, a steering angle sensor (not

[0123] 303774181.1 pictured) can be included to further increase the system 200 reliability. Additionally, the set of sensor can comprise at least one of a throttle actuator position sensor, or a motor speed sensor, or an impeller shaft speed sensor, or a driveshaft speed sensor . According to an embodiment, the IMU 250 can also be configured to read magnetic field to act as a compass. This can be useful, since the watercraft can "turn on itself" (contrarily to most land vehicles). This motion cannot be captured by GPS, as GPS needs a trace over a distance ("breadcrumbs") to compute a heading of a watercraft.

[0124]

[0105] According to an embodiment, the combination of GPS 240 and an IMU 250 provides more precise location data compared to using only one sensor. By merging the data from both sensors, the system 200 can compensate for GPS signal weaknesses or interruptions, resulting in improved overall accuracy.

[0125]

[0106] The integration of GPS 240 and IMU 250 improves the robustness of the navigation system. In environments where GPS signals are weak or unavailable, the IMU data can help maintain a stable reference frame for the system 200. Conversely, in situations where GPS signals are strong but prone to interference, such as during ionospheric disturbances for example, the IMU data can help filter out noise and provide more reliable position information.

[0126]

[0107] By using an IMU 250 to estimate small movements between GPS fixes, the system 200 can reduce the frequency of GPS signal acquisitions, thereby increasing the frequency at which a position can be received, using for example methods such as Kalman filtering,.

[0127]

[0108] The use of an IMU 250 allows for faster response times in navigation systems. Since the IMU 250 measures angular velocity and linear acceleration continuously, it can provide real-time data to the system 200, enabling quicker adjustments to changes in position or orientation. This is especially useful in applications where rapid response is necessary, such as watercraft 10.

[0128]

[0109] According to an embodiment, the watercraft 10 can comprise a plurality of modules. For example, the watercraft 10 can comprise a notification module 260. This module 260 is configured to alert the user, and, for example, to generate notifications, as described hereafter. As described above, the watercraft 20 comprises a screen 231 configured to display information, for example in real-time. Displaying real-time information on a screen 231

[0129] 303774181.1 provides the user with instant access to critical data, enabling them to make informed decisions quickly. This can be particularly useful in dynamic environments where conditions can change rapidly, such as during a ride on a watercraft 10. By having real-time data readily available, the user can respond to situations more efficiently and reduce potential risks.

[0130]

[0110] Moreover, the screen 231 may be connected to various sensors and systems on the watercraft 10, providing real-time data from multiple sources. This can comprise ride data, weather information, navigation data, and system status indicators.

[0131]

[0111] The screen 231 can also serve as an interface for users to interact with the system 200. For example, the screen 231 can be a touchscreen and can have a graphical user interface (GUI) display that supports multiple input methods, including voice commands, gestures, and keyboard inputs. This versatility allows users to interact with the system 200 in various ways, depending on their preferences and the specific application requirements. The visual notification aspect offers a clear and distinct visual indication of an event or condition. This can be achieved through various means such as flashing lights, changing colors, or graphical representations, . Visual notifications provide a quick and easily noticeable way to convey information at a glance.

[0132]

[0112] According to an embodiment, the watercraft 10 can comprise at least one speaker configured for generating audible notifications to a user. For example, the speaker is integrated within the system 200 design and can produce various types of audible alerts or messages. The use of a speaker in the system 200 enables real-time communication with the user through audio notifications, enhancing overall functionality and usability. Additionally, the speaker maybe adjustable to different volume levels, ensuring that the audio notifications can be heard clearly under various environmental conditions. Moreover, a multiple or two speakers system with a stereo sound can be implemented to give an audio queue a directionality aspect. For example, the sound system can be configured to generate sound according to the direction of an incoming safety warning. The audio notification can also comprise sound feedback.

[0133]

[0113] According to an embodiment, the watercraft 10 can comprise a vibration device. This device is configured to generate haptic notifications for a user. For example, the use of haptic feedback in place of or in addition to audible and visual notifications can help reduce distractions and improve user experience, particularly in noisy environments or when the

[0134] 303774181.1 user is unable to look at the device's screen 231. For example, the vibration device can be located in the handles, in the seat or in the feet rests to provide haptic notifications or feedback to the user.

[0135]

[0114] Furthermore, the vibration device can be designed to provide different types of haptic feedback, such as varying intensity, duration, or pattern, to distinguish between different notifications and enhance user awareness. This allows also to distinguish from other vibrations coming from e.g. environment (waves, ripples, etc.) and e.g. watercraft (engine, propeller, etc.).

[0136]

[0115] According to an embodiment, the pre-trained neural network is configured to detect at least one object 20 within a predetermined area, the predetermined area being located in front of the watercraft. To do so, the pre-trained neural network is configured to use data collected by the perception-based sensor 210. According to an embodiment, the predetermined area can be the field of view of the perception-based sensor 210. The network can act in two steps : a) In response to the detection at least one object, creating a "zone of interest" (a subarea of the full sensor view, i.e. of the predetermined area). This allows a fast response as well as the use of an inexpensive computational network, helping with the control of energy consumption and computational time; and b) An "in depth" analysis of the "zone of interest" by at least one neural network, allowing for greater detail detection. This neural network can be the same than in the previous step or it can be a different one.

[0137]

[0116] As further discussed, hereafter, according to an embodiment, this pre-trained neural network could be trained on real dataset using data from a GPS module 240 and from a LiDAR module 250 as well as on synthetic dataset using simulated data. The dataset used for the training may be labelled, i.e. annotated, with key attributes such as size and / or distance of objects. This dual training approach enables the pre-trained neural network to learn patterns and relationships between sensor data and object trajectories in various scenarios, thereby enhancing its ability to detect objects and predict their trajectories with high accuracy.

[0138]

[0117] As illustrated by figures 6 and 7, real-world data can be obtained from experiments conducted on the watercraft 10, where the GPS and LiDAR modules 240, 250 provide accurate

[0139] 303774181.1 spatial information about the surroundings. This data comprises various environmental conditions such as fog, snow, darkness, and different types of obstacles. The real-world data is used to train the neural network to learn patterns and relationships between sensor data and object trajectories in various scenarios. In figure 6, x-axis represents the distance traveled by the detected object 20 and the y-axis is the distance between the watercraft 10 and the detected object 20. The jagged line is the estimated data obtained from the neural network while the other one is the GPS data. In figure 7, the distance between the watercraft 10 and the detected object 20 is calculated from the data acquired by the perception-based sensor.

[0140]

[0118] Synthetic data, on the other hand, can be generated through simulation to mimic situations that cannot be easily realized in real life for security reasons. This comprises scenarios such as extreme weather conditions, complex obstacle courses, or rare events like close calls with other watercraft, swimmer, animal, or obstacles. For purposes of this application, a "close call” is a situation where a collision almost happens. The synthetic data is used to augment the real-world data and provide the neural network with a more comprehensive understanding of potential hazards and obstacles.

[0141]

[0119] According to an embodiment, the pre-trained neural network is trained using machine learning techniques, for example convolutional neural networks (CNNs) and / or recurrent neural networks (RNNs). The CNNs can be used for object detection and classification, while the RNNs can be used to predict object trajectories. According to an embodiment, the training process involves a combination of supervised and unsupervised learning, where the network is trained on labeled data to learn patterns and relationships between sensor data and object trajectories.

[0142]

[0120] According to an embodiment, and as represented by figures 3, 6, 7, 13 to 17, the present technology refers to a method 100 for perception-based trajectory determination for watercraft 10. This method 100, as described hereafter, utilizes the pre-trained neural network to detect objects 20 within a designated observation area in front of the watercraft 10. According to this technology, real-time data from the perception-based sensor 210 is fed into the pre-trained neural network. When an object 20 is detected, relevant information is extracted and used to calculate predicted trajectories for both the detected object 20 and the watercraft 10. The method 100 computes information using this acquired data and data from

[0143] 303774181.1 sensors on the watercraft 10. This computed information comprises the predicted trajectory 21 of the detected object 20 and the predicted trajectory 11 of the watercraft 10. Then, a situation is assessed by the present technology and measures are taken based on the risk involved. Indeed, based on the calculated information, safety actions can be executed according to predefined rules. These actions can include one or more of warning the user of potential collisions, adjusting the watercraft's heading, reducing motor power output, and / or activating automatic braking of the watercraft 10.

[0144]

[0121] According to an embodiment, and with reference to figure 3, the method 100 comprises: a) accessing 110 the pre-trained neural network, using the neural network module 220, to detect an object 20 within a predetermined area of observation, the predetermined area of observation being located around the watercraft 10, and in some embodiments in front of the watercraft 10, forming a field of observation; b) acquiring 120, by the processing module 230, data from the perception-based sensor 210 mounted on the watercraft 10 and feeding this acquired data to the pre-trained neural network, in real-time; c) detecting 130, by the neural network module 220, an object 20 located in the predetermined area of observation based on the acquired data from the perceptionbased sensor 210; d) extracting 140 a first set of information regarding the detected object 20, the first set of information being configured to be used to determine a predicted trajectory 21 of the detected object 11, the first set of information comprising the relative distance between the detected object and the watercraft, the detected object position based on the coordinates of the detected object within a field of view of the perception-based sensor; and the relative bearing between the detected object and the watercraft. According to an embodiment, the first set of information can also comprise an estimation of the detected object attitude comprising at least one of an estimation of the pitch of the detected object 20 or the roll of the detected object 20, or the yaw of the detected object 20;

[0145] 303774181.1 e) acquiring 150 a second set of information about the watercraft 10 from the set of sensors of the watercraft 10; f) computing 160 a third set of information using the first set of information and the second set of information, the third set of information comprising a predicted trajectory 21 of the detected object 20, a predicted trajectory 11 of the watercraft 10 and a relative speed between the detected object 20 and the watercraft 10; g) Determining 170 the closest distance, also called the "close call distance” CCD between the trajectories as a function of time between the detected object 20 and the watercraft 10, based on the predicted trajectory 21 of the detected object 20, the predictive trajectory 11 of the watercraft 10 and the relative speed between the detected object 20 and the watercraft 10; According to an embodiment, the closest distance CCD corresponds to the point in which the detected object and the watercraft are temporally closest. This is an input that can be used into a variety of tables to determine the severity / probability, as discussed hereafter. According to an embodiment, the closest distance CCD is used in conjunction with predetermined assessment tables to generate a normalized event value. The normalized event value may be configured to quantify the collision probability and / or severity based on the computed closest distance and / or additional parameters. The assessment tables can comprise lookup tables and / or algorithmic functions that correlate the closest distance CCD with predefined event thresholds and / or predetermined margins. h) in response to the determination of the closest distance, executing 180 a predetermined action based on the computed third set of information. According to an embodiment, the predetermined actions are executed progressively or with predetermined limits based on the normalized event value. The progressive execution may comprise implementing incremental measures, corresponding, for example, to increasing event levels. The predetermined actions with predetermined limits may comprise activating specific measures when the normalized event value exceeds defined thresholds.

[0146]

[0122] According to an embodiment, the method 100 further comprises a step of monitoring the steering capability of the watercraft 10 using a steering angle sensor. The processing

[0147] 303774181.1 module 230 may be configured to determine whether the operator is actively changing the watercraft trajectory or whether the steering mechanism has reached its operational end stops. For example, when the steering angle sensor indicates trajectory modification by the operator or detection of steering end stops, the processing module 230 selectively restores previously reduced throttle power. This restoration maintains the watercraft's maneuverability while reducing collision avoidance capabilities.

[0148]

[0123] According to an embodiment, the management of the throttle can operate in conjunction with the steering angle sensor to provide adaptive power control. The processing module 230 may continuously monitor steering input and may dynamically adjust the throttle restoration based on the detected steering activity. The throttle restoration can be proportional to the steering angle deviation and the current normalized event value, ensuring balance between collision avoidance and watercraft control.

[0149]

[0124] According to an embodiment, this method 100 further comprises a step of providing the user of the watercraft 10 with real-time information about the detected object 20. According to an embodiment, the object 20 can be taken among various types of objects. These objects comprise an animal, another watercraft, a swimmer, or a natural obstacle.

[0150]

[0125] According to an example, and as illustrated by figure 13 the screen 231 is configured to display a schematical representation of the position of the detected object 20 regarding the position of the watercraft 10. Displaying a schematical representation of the object's position relative to the watercraft 10 on the screen 231 allows the operator to quickly and easily understand the spatial relationship between the watercraft 10 and the detected object 20. This information is useful for navigation. By providing a clear visual representation of the detected object's position, the system 200 enables operators to make informed decisions regarding the watercraft's trajectory and speed. According to another embodiment, the screen 231 can be configured to display a live feed from the perception-based sensor, with optional color, messages, drawings, etc. on top of the live feed, for example.

[0151]

[0126] According to an embodiment, this method 100 utilizes the pre-trained neural network to detect objects 20. In real-time, data from the perception-based sensor 210 mounted on the watercraft 10 are acquired and fed into the pre-trained neural network. When an object 20 is detected, a first set of information regarding that object 20 is extracted. This first set of information can comprise the object's position, size, shape, attitude, and

[0152] 303774181.1 velocity, and is used to determine a predicted trajectory 21 of the detected object 20, see for example figure 17, as well as the relative speed between the detected object 20 and the watercraft 10. According to an embodiment, the second set of information about the watercraft 10 is acquired from the set of sensors such as a GPS 240 incorporated into a navigation system for example, an IMU 250, or steering angle sensor.

[0153]

[0127] According to an embodiment, and as represented by figure 17, the system 200 integrates trajectory estimation and probability analysis to anticipate and mitigate potential collisions. It utilizes estimated trajectories of both the watercraft 10 (HV) and the detected object 20 (RV) to ascertain the closest distance between the trajectories as a function of time called Close Call Distance (CCD). Afterwards two values are calculated the first being the time to the closest distance which is the amount of time until the CCD point along the watercraft’s trajectory and doesn’t necessarily indicate that a collision will occur, the second being a direct distance which is the amount of distance the watercraft 10 must take along the trajectory before the closest distance is hit. Based on these values as well as both detected object’s dynamic and watercraft’s dynamic the system 200 can be configured to impose a progressive throttle restriction or other actions, thereby helping to reduce the probability of a collision. For example, the throttle limitation can provide information to the driver to adjust the watercraft's trajectory in real-time, providing an opportunity to reduce the probability of a collision by aligning the watercraft with a safer path. This proactive approach ensures that the driver is alerted to potential hazards and can take corrective action to prevent accidents. In an embodiment a probability is calculated based on the closest distance CCD and on the direct distance. According to an embodiment, the present technology uses calibrated tables, like for example an assessment table. These tables are configured to provide, for example, a probability based on the distance between the detected object 20 and the watercraft 10 at their closest point temporally. For example, a close call of 0 would be a probability of 1. This is for example a factor that is applied to a reduction that is based on a relative speed and a relative velocity. Indeed, according to an embodiment, a normalized event value can be calculated or generated using the closest distance CCD and at least one predetermined event assessment table.

[0154]

[0128] According to an embodiment, and as illustrated by figure 18, the present technology can use a sub-method configured to calculate the required limitation using a plurality of tables. For example, in this figure 18, two main tables are used. These tables can also be or

[0155] 303774181.1 comprise vectors. The first table, called table 1, relate to the relative distance between the watercraft 10 and the detected object 20. The second table or vector, called vector 2 or table 2, is related to the closest distance CCD. According to an embodiment, this sub-method comprises subtracting the value of the first table to 1 and then multiplying the result by a factor from table 2; then, all of this is subtracted from 1 again to give a value which is the maximum available throttle. In more detail, the table 1 associates some parameters’ values with a calculated relative distance, and the table 2 associates some parameters’ values with the calculated closest distance CCD. The output of table 2 is, for example, a probability.

[0156]

[0129] According to an embodiment, and as illustrated by Figure 19, the present technology can use a filtering sub-method configured to filter some sensor data. According to an example, this filtering sub-method comprises values for rising and falling. The falling is related to the fact that the watercraft 10 should be able to reduce its speed or acceleration quickly without exceeding a predetermined threshold. The rising relates to the fact that a sudden return of throttle should avoid to exceed another threshold.

[0157]

[0130] According to an embodiment, the rising behavior implements a gradual throttle restoration mechanism that prevents sudden acceleration jerks when throttle limitations are removed. When the operator maintains maximum throttle input during a limitation event, the present technology is configured to monitor the throttle position and to withhold throttle restoration until the operator reduces throttle input below maximum levels. Once throttle input drops below 100%, the present technology is configured to restore available throttle from the limited percentage back to the operator's desired throttle level at a controlled rate. This rising mechanism prevents the watercraft 10 from experiencing abrupt acceleration transitions that would occur if throttle were instantly restored from a limited state to full operator demand.

[0158]

[0131] According to an embodiment, the falling behavior implements a rapid but controlled throttle reduction mechanism to ensure safe deceleration without causing abrupt vehicle dynamics that could destabilize the watercraft 10 or create hazardous conditions for the operator. When throttle limitation is applied, the present technology is configured to reduce power output quickly to reduce the probability of collision scenarios while maintaining a gradual transition rate that prevents the watercraft 10 from experiencing sudden deceleration from maximum throttle to zero throttle, for example within fractions of seconds.

[0159] 303774181.1 This controlled falling response ensures the operator maintains vessel control during emergency limitation events while providing sufficient response time to react to the throttle reduction.

[0160]

[0132] According to an embodiment, the first set of information comprises: a) Relative distance between the object 20 and the watercraft 10, i.e., the measurement of the spatial separation between the object 20 and the watercraft 10. The determination of this parameter is useful for assessing the proximity of the object 20 to the watercraft 10, see for example figures 6 and 7, which can be useful in various applications such as collision reduction or object recognition. b) Object 20 position based on its coordinates within the perception-based sensor's field of view 211, 212: This involves determining the spatial location of the object 20 within the perception-based sensor's field of view 211, 212. By calculating the object's coordinates, the system 200 can track and monitor the movement of the detected object 20 relative to the watercraft 10, and can predict its future positions, therefore its trajectory 21. c) Relative bearing between the object 20 and the watercraft 10: This represents the angular relationship between the object 20 and the watercraft 10, see for example figure 16. The determination of the relative bearing is useful for identifying the direction of the object 20 with respect to the watercraft 10, which can be useful in navigation or for reducing the probability of collision situations. d) Object attitude estimation: This involves estimating the orientation or posture of the detected object 20. By determining the object's attitude, the system 200 can gain insights into the object's position and motion in three dimensions. According to an embodiment, this estimation can require using a pose neural network configured to convert a position into an angular position or to directly estimate the angular position of the detected object 20. According to an embodiment, some image processing can be implemented to compensate some spatial distortions, for example.

[0161] 303774181.1

[0133] The use of a first set of information comprising relative distance, object position, relative bearing, and object attitude estimation allows for improved accuracy and reliability in the detection, tracking, and analysis of objects 20 in relation to the watercraft 10.

[0162]

[0134] According to an embodiment, the present technology can also utilize other onboard sensors such as a steering sensor, acceleration sensors, and a gyroscope to increase the accuracy of the avoidance calculations. These sensors provide additional information about the watercraft's attitude and velocity, which can be used to refine the collision prediction and response.

[0163]

[0135] According to an embodiment, the present technology can use sensor fusion techniques allowing for combining data from multiple sensors to improve the accuracy of the attitude estimation.

[0164]

[0136] For example, data from an accelerometer, gyroscope, and magnetometer can be fused using a Kalman filter or other fusion algorithm to provide more reliable attitude estimates than relying on any single sensor alone.

[0165]

[0137] Furthermore, the method 100 may comprise calibrating the sensors used for attitude estimation to ensure accurate readings. This calibration process may involve periodic adjustments based on known reference points or using self- calibration techniques.

[0166]

[0138] According to an example, the second set of information comprises: a) Position via the GPS module 240 integrated into a navigation system of the watercraft 10; b) Attitude of the watercraft 10 via the IMU module 250, a compass sensor, or vehicle dynamics model.

[0167]

[0139] The integration of a GPS module 240 into the navigation system allows for precise determination of the watercraft's position in real-time, enabling efficient navigation and location tracking. It enhances safety by providing accurate information on the watercraft's position, which is helpful for reducing the probability of collisions with other vessels or obstacles.

[0168] 303774181.1

[0140] The IMU module 250 measures the angular speed and linear accelerations of the watercraft to determine its attitude (pitch, roll, and yaw angles). This data is helpful for maintaining stability and controlling the watercraft, as well as for predicting its trajectory 11.

[0169]

[0141] The compass sensor provides heading information, which is useful for navigation and the computing of the predicted trajectory 11.

[0170]

[0142] Additionally, the vehicle dynamics model helps analyze the watercraft's behavior under various conditions, allowing for optimized performance and improved safety.

[0171]

[0143] According to an example, the third set of information comprises: a) Relative speed between the watercraft 10 and the detected object 20 using the first set of information and the second set of information; b) Predicted trajectories 11, 21 for both the watercraft 10 and the detected object 20.

[0172]

[0144] The inclusion of relative speed information between the watercraft 10 and the detected object 20 in the third set allows for a more accurate assessment of the proximity and potential collision risk between them.

[0173]

[0145] The addition of predicted trajectories 11, 21 for both the watercraft 10 and the detected object 20 provides a more comprehensive understanding of their movements. This information enables the system 200 to anticipate potential collisions or obstacles, allowing the watercraft operator to take preventative measures before an incident occurs.

[0174]

[0146] According to an embodiment, the method 100 comprises calculating the relative distance between the watercraft 10 and the detected object 20 using at least one of computer vision or depth estimation algorithms. The computer vision algorithm can process visual data from the perception-based sensor 210 to identify and analyze features of the scene, enabling the determination of distances based on image analysis. The depth estimation algorithm can utilize techniques such as stereo vision, time-of-flight, or structured light to measure the distance between objects in a scene by analyzing the intensity or phase differences of reflected light. By employing these algorithms or others, the method 100 can accurately determine the relative distances between the watercraft 10 and objects 20, providing valuable information for collision avoidance.

[0175] 303774181.1

[0147] According to an embodiment, the method comprises calculating the relative bearing between the detected object 20 and the watercraft 10. This is achieved by utilizing the detected object's position and the watercraft's heading. According to an embodiment, the detected object's position refers to its relative coordinates according to the watercraft 10. The watercraft's heading, on the other hand, represents its direction of travel in degrees relative to true north, for example.

[0176]

[0148] According to an embodiment, the calculation of the relative bearing is performed using trigonometric functions such as sine and cosine. The difference between the detected object's azimuth (the direction it lies from the watercraft 20) and the watercraft's heading is calculated to determine the relative bearing.

[0177]

[0149] According to an embodiment, as previously mentioned, estimating the attitude of the detected object 20 comprises determining the pitch, roll, and yaw of the object 20. The pitch estimation refers to the angular displacement of the detected object 20 around a lateral axis. This measurement is useful for assessing the detected object's movement in the fore-and-aft direction. The roll estimation pertains to the angular displacement of the detected object 20 around a longitudinal axis. This measurement helps determine the detected object's lateral tilt or deviation from the vertical position. Lastly, the yaw estimation represents the angular displacement of the detected object 20 around a vertical axis. This measurement is useful for assessing the detected object's heading in a horizontal plane. This allows for comprehensive attitude estimation of the detected object 20, providing valuable information for predicting its trajectory 21.

[0178]

[0150] According to an embodiment, estimating the detected object attitude uses machine learning algorithms. For example, the machine learning algorithm is pre- trained on historical data to identify patterns and make predictions about the detected object's attitude based on input data from at least the perception-based sensor. The algorithm maybe a neural network, support vector machine, or other suitable machine learning model.

[0179]

[0151] According to an embodiment, the determination of the relative velocity of the detected object 20 can be achieved through various techniques, such as tracking the position of the object over time and calculating the change in position between successive measurements. The velocity data can be obtained using sensors or other measurement devices. Incorporating the relative velocity information into the method 100 allows for more

[0180] 303774181.1 accurate analysis and interpretation of the data related to the detected object 20. For instance, this information can be used to identify patterns or trends in the motion of the detected object 20, which may have practical applications in predicting the future trajectory of the detected object 20.

[0181]

[0152] According to an embodiment, the third set of information comprises at least two distinct subsets.

[0182]

[0153] The first subset of the third set of information can comprise the relative speed between the watercraft 10 and the detected object 20. This information can be derived from the first and second sets of data.

[0183]

[0154] Moreover, the second subset of the third set of information pertains to predicted trajectories 11, 21 for both the watercraft 10 and the detected object 20. These predicted paths 11, 21 can be calculated based on the past and or current velocities and positions of each entity, as well as external factors such as wind or water currents, for example. By incorporating the relative speed and predicted trajectory information, the method 100 can provide more comprehensive situational awareness for the watercraft user. This data can be utilized in various applications, including collision avoidance systems, navigation assistance, and environmental monitoring.

[0184]

[0155] According to an embodiment, the method 100 may comprise determining the nature of the detected object 20 as part of the information about the detected object 20. For example, this information is obtained through at least the perception-based sensor 210 or other data collection means. Incorporating the nature of the detected object 20 into the information allows for more specific identification and classification of the detected object 20 and can help to refine the predicted trajectory 21. By having access to more comprehensive information about an object, the system 200 can make better decisions based on that data. In some embodiments, the nature of the detected object may be determined using various techniques such as spectral analysis, machine learning algorithms, or other relevant methods. Additionally, it is interesting to note that the method is not limited to identifying only the nature of an object; it can also comprise other types of information about the object, such as its shape, size, or color. However, including the nature of the detected object in the information enhances the quality of the predicted trajectory 21.

[0185] 303774181.1

[0156] According to an embodiment, the present technology is configured to provide the user with real-time information in a manner that enhances situational awareness. For example, this can comprise displaying the information on the screen 231, which is readily visible and accessible to the user. The use of a screen 231 enables the presentation of complex data in an easily digestible format, such as graphs, charts, or animations. Furthermore, the display may be customizable, allowing the user to configure it according to their preferences or specific ride conditions. For example, the display may prioritize certain information based on the phase of the ride or the type of watercraft 10 being operated or the type of detected object 20.

[0186]

[0157] According to an embodiment, the method 100 further comprises, before computing the third set of information, keeping a record of the timestamp for the acquired data. Recording the timestamp prior to computing the third set of information allows for accurate tracking and analysis of when each data processing step occurs. This feature is particularly useful in applications where real-time or historical data analysis is required.

[0187]

[0158] In some embodiments, the recording of timestamps may also comprise metadata such as the processing node or thread responsible for each step, allowing for further analysis of parallel processing systems.

[0188]

[0159] According to an embodiment, the step of determining the closest distance between the detected object 20 and the watercraft 10 involves analyzing the predicted trajectories 11, 21 of both entities 10 and 20. This is done by comparing the future positions of the detected object 21 and the watercraft 10, taking into account their respective velocities and any external factors that may influence their motion.

[0189]

[0160] To determine if an impact will occur or to determine the closest distance CCD, the processing module 230 computes the closest distance between the detected object 20 and the watercraft 10 according to the two predicted trajectories 11, 21 and to their respective speeds and / or acceleration. If the closest distance lies within a predetermined time or distance threshold, it is determined that a close call distance exists. This information is then used to trigger a safety action, such as alerting the watercraft user or executing an evasive maneuver. For example, to determine if a risk of impact is possible, the processing module 230 computes the closest distance between the two trajectories 11, 21 which takes into account their respective speeds, positions and sometimes accelerations. If the closest

[0190] 303774181.1 distance is within a predetermined time or distance (which can be adjusted using a factor K based on the watercraft’s dynamics) the information is used to trigger a safety action, such as alerting the watercraft user or executing an evasive maneuver.

[0191]

[0161] According to an embodiment, the system 200 implements a multi-tiered risk assessment framework wherein collision risk is evaluated as a continuous variable. The processing module 230 may be configured to calculate a probability value based on the closest distance CCD, wherein the risk level increases progressively as the closest distance decreases and / or the time to closest distance diminishes. The safety actions may be implemented progressively according to predetermined threshold levels. For example, when the time closest distance falls below a first predetermined time threshold or the direct distance falls below a first predetermined distance threshold, the system 200 can be configured to initiate a warning notification to the watercraft user. As the risk escalates, when the time closest distance or direct distance falls below a second predetermined threshold, the system 200 may reduce motor power output below a predetermined limit. At the highest risk level, when the time closest distance or direct distance falls below a third predetermined threshold, the system 200 may execute automatic braking of the watercraft using reverse gate activation, for example. The predetermined thresholds can be dynamically adjusted using factor K based on the watercraft's dynamics, comprising velocity, acceleration, and / or maneuverability characteristics. This progressive throttle restriction approach provides the driver with real-time information to adjust the watercraft's trajectory, thereby creating opportunities to reduce the probability of collision by aligning the watercraft with a safer path while maintaining proportional response to the assessed risk level.

[0192]

[0162] The determination of a closest distance, a time to the closest distance and a direct distance is based on a combination of factors, including the predicted trajectories 11, 21 of the detected object 20 and the watercraft 10, as well as their relative speed and any external influences. By analyzing these factors, the system 200 can provide accurate and timely warnings to prevent potential collisions. According to an embodiment, the time to the closest distance corresponds to the time to reach a point of the predicted trajectory 11 of the watercraft 10 corresponding to the point of the closest distance.

[0193]

[0163] According to an embodiment, depending on the perception-based sensor 210 being used and the speed of the watercraft 10, the area surveyed around or in front of the watercraft

[0194] 303774181.1 10 may vary. Most sensors limit the area to a triangle-shaped region on either side of the watercraft 10 centerline. However, for faster watercraft 10, a smaller angle can be used to allow for collision avoidance maneuvers. This area surveyed can be dynamically updated based on the vehicle's speed and user steering angle input. According to an embodiment, this triangle shaped region can also be affected by the steering angle, allowing the triangle to deviate by a factor to the left or right depending on steering direction; giving information about driver intent, for example.

[0195]

[0164] According to an embodiment, if a collision possibility is calculated along the watercraft's trajectory 11, the processing module 230 may limit the acceleration or speed of the watercraft 10 as a warning to the driver. To prevent confusion, the present technology, as illustrated in figure 13, can be configured to display graphical representations of the watercraft 10, the triangle of the perception-based sensor 210, and the detected object 20, i.e., the potential colliding object. An audio warning can also be provided depending on the distance between the watercraft 10 and the detected object 20.

[0196]

[0165] According to an embodiment, based on the computed third set of information, safety actions are indeed executed according to a predetermined set of safety rules. These safety actions can comprise several actions according to the situation, as explained above.

[0197]

[0166] According to an embodiment, once the pre-trained neural network analyzes the data from the perception-based sensor 210, it determines the distance and relative bearing between the watercraft 10 and the detected object 20. The attitude (pitch, roll, and yaw) of the watercraft 10 and the object 20 can also be taken into account to increase accuracy. Furthermore, a tracking algorithm can be put in place to follow objects once detected and determine trajectory 22. Which can be further used as described to calculate a closest distance CCD, a time to the closest distance and a direct distance.

[0198]

[0167] For example, an onboard computer, also called the processing module 230, processes this data to calculate the trajectories 11, 21 of both the watercraft 10 and the object 20, and determines if the closest distance. The driver of the watercraft 10 is then notified in advance of a potential collision through any combination of audio or visual or haptic means. If a time to the closest distance or a direct distance before reaching the closest distance is less than a predetermined amount, the present technology can automatically reduce the watercraft's speed to help reducing the probability of a collision while maintaining turning capabilities,

[0199] 303774181.1 i.e. while maintaining a minimum of throttle to allow for turning due to the nature of a jet boat, for example.

[0200]

[0168] According to an embodiment, the present technology can also employ Kalman filtering techniques to calculate the trajectories of the watercraft 10 and of the detected object 20. This estimation can rely on GPS data for example and can be improved by the fusion of additional sensors such as steering angle and IMU. Indeed, according to an embodiment, the prediction of trajectories uses Kalman filtering. In particular, the method involves initializing a state vector and covariance matrix based on available data. Subsequently, the method applies recursive calculations to update the state vector and covariance matrix based on new measurements and process noise. The predicted future trajectory is then obtained from the updated state vector. Using Kalman filtering in this method allows for efficient estimation of the system's state and prediction of future trajectories despite the presence of measurement and process noise. Additionally, the recursive nature of the algorithm enables real-time implementation and adaptation to changing conditions. The use of Kalman filtering provides a robust solution for predicting future trajectories by taking into account both measurement errors and process noise. Indeed, in real-world applications, data is often subjected to noise or measurement errors. The use of the Kalman filtering method in predicting trajectories provides robustness against such disturbances. The filtering algorithm utilizes a mathematical model that represents the relationship between the system's state and measurements, enabling it to reduce noise and provide reliable predictions even under challenging conditions.

[0201]

[0169] Systems with changing dynamics, like watercraft 10 or objects 20 on water, require methods capable of adapting to these variations. Kalman filtering offers this adaptability by recursively updating its estimates based on new data. This allows the present technology to account for changes in system behavior, making it an ideal choice for predicting trajectories in complex and evolving environments.

[0202]

[0170] Compared to other advanced prediction methods, Kalman filtering offers reduced computational complexity. This makes it an attractive choice for the present technology where computational resources can be limited on a watercraft 10.

[0203] 303774181.1

[0171] According to an embodiment, using historical data, information such as relative speed and trajectory can be computed for both the object 20 and the watercraft 10. Advanced algorithms and filters may be required for some of these calculations.

[0204]

[0172] According to an embodiment, the third set of information is used to calculate the time to the closest distance between the watercraft 10 and the detected objects 20 and to determine the necessary throttle limitation to reduce the chances of a collision. This information can also be used to calculate the direct distance between the actual position of the watercraft 10 and the closest distance, i.e. the close call point.

[0205]

[0173] The time to the closest distance is a parameter in collision avoidance systems. It represents the time remaining until an object or vehicle has a close call with another object, vehicle, or surface . In the context of the present technology, the time to the closest distance is determined based on the predicted trajectory 21 of the detected object 20, the velocity of the detected object 20 and the current position and velocity of the watercraft 10. The time to the closest distance can be calculated using various methods, including dynamic equations that take into account the relative motion between the watercraft 10 and the detected object 20. The result is a time value that indicates how long it will take for the two objects to reach the CCD point , assuming their current trajectories continue unchanged as well as their respective speeds, accelerations, headings, etc. This information can be used to trigger warnings or alerts to the user, adjust the watercraft's heading or speed, or even initiate emergency braking procedures.

[0206]

[0174] In some embodiments, the time to the closest distance may be calculated in real-time using sensor data from various sources, such as GPS 240, IMU 250, and perception-based sensor 210. The processing module 230 can then use this information to determine the time to the closest distance and trigger appropriate responses based on predefined thresholds and rules.

[0207]

[0175] The "direct distance” refers to the spatial separation between the actual position of the watercraft 10 and the predicted point of closest distance between the watercraft 10 and the detected object 20. This value represents the remaining distance that can be traversed by the watercraft 10 before getting to the CCD point, taking into account both objects current velocity, position, acceleration and therefore trajectory.

[0208] 303774181.1

[0176] In other words, the direct distance is a measure of how far away the watercraft 10 is from the point where it gets close to the detected object 20, assuming no changes in course or speed. This value is typically calculated using kinematic equations that take into account the relative motion between the watercraft 10 and the detected object 20.

[0209]

[0177] The direct distance value is a useful input for the collision avoidance system, as it helps determine when to initiate safety actions such as warning the user, adjusting the heading, reducing motor power output, or automatically braking the watercraft 10.

[0210]

[0178] According to an embodiment, the method 100 further comprises, in response to the determination of the closest distance, determining at least one of: a) the direct distance between the actual position of the watercraft 10 and the CCD point; or b) the time to the closest distance.

[0211]

[0179] According to an embodiment, the safety actions further comprise: a) warning the user via a notification in response to at least one of the time to the closest distance being smaller than a predetermined first time threshold or the direct distance being smaller than a first predetermined distance threshold; b) reducing motor power output below a predetermined limit in response to at least one of the time to the closest distance being smaller than a predetermined second time threshold or the direct distance being smaller than a second predetermined threshold; and c) automatic braking of the watercraft 10 using a reverse gate 86 in response to at least one of the time to the closest distance being smaller than a predetermined third time threshold or the direct distance being smaller than a third predetermined threshold. d) It is contemplated that in some embodiments, the actions could additionally include adjusting the watercraft's heading based on the detected object's predicted trajectory 21 to maintain a safe distance in response to at least one of the time to the closest distance being smaller than a predetermined fourth time threshold or the direct distance being smaller than a fourth predetermined threshold;

[0212]

[0180] According to an embodiment, the first distance threshold is higher than the second distance threshold. The second distance threshold is also higher than the third distance threshold. According to an embodiment, the third distance threshold is higher, equal or lower than the fourth distance threshold.

[0213] 303774181.1

[0181] For example, the first distance threshold is equal or higher than 50m the second distance threshold is comprised between 40m and 50m, the third distance threshold is comprised between 30m and 40m: According to an embodiment, the fourth distance threshold is comprised between 20m and 40m.

[0214]

[0182] According to an embodiment, the values or ranges of these distance thresholds can be based on calibrated maps or calibrated tables, also called lookup tables. These maps or tables can be configured to provide an output according to an input, this input being mono or multi-dimensional, like distance, speed, acceleration, etc....

[0215]

[0183] The use of multiple distance thresholds in this manner allows for a more granular control over the distance parameters in the method. By setting the first threshold as the highest, followed by successively lower thresholds, the method can distinguish between different distances and respond accordingly. This hierarchical arrangement of thresholds may lead to improved accuracy and efficiency in the implementation of the method.

[0216]

[0184] According to an embodiment, the first time threshold is higher than a second time threshold, the second time threshold is higher than the third time threshold. According to an embodiment, the third time threshold is higher, equal or lower than the fourth time threshold.

[0217]

[0185] For example, the first time threshold is equal or higher than 120 seconds, the second time threshold is comprised between 60 seconds and 120 seconds, the third time threshold is comprised between 30 seconds and 60 seconds. According to an embodiment, the fourth time threshold is comprised between 15 seconds and 30 seconds.

[0218]

[0186] As for the distance thresholds, these time thresholds can be based on lookup tables, for example.

[0219]

[0187] Setting different actions according to different time or distance thresholds allows one to adapt the safety action.

[0220]

[0188] According to an embodiment, the method 100 can comprise adapting the set of safety rules. This adaptation can be carried out in response to different types of environmental conditions. The environmental conditions may encompass various factors such as temperature, humidity, pressure, or velocity of the wind. The method 100 can respond to these conditions by modifying the safety rules accordingly. For instance, under extreme wind,

[0221] 303774181.1 certain safety rules may need to be tuned to consider the impact of the wind. The adaptation of safety rules can be achieved through various means, such as using sensors to monitor environmental conditions and adjusting the rules in real-time or periodically based on historical data. The method 100 may also comprise notifying the user of any changes to the safety rules and providing them with appropriate instructions.

[0222]

[0189] This adaptation feature enhances the overall safety and reliability of the system 200 by ensuring that safety rules are tailored to specific environmental conditions, thereby reducing the risk of accidents or collisions. It also allows for greater flexibility in implementing safety rules, as different environments may require distinct safety measures.

[0223]

[0190] According to an embodiment, the method 100 can comprise adapting the first time threshold or the second time threshold or the third time threshold or the fourth time threshold based on different types of environmental conditions.

[0224]

[0191] The environmental conditions can be sensed using appropriate sensors and the data obtained is used to determine the need for adjusting the time thresholds.

[0225]

[0192] Adapting the time thresholds in response to environmental conditions can improve the efficiency and accuracy of the method 100. For instance, in harsh environments where quick responses are required, reducing the first time threshold can be beneficial. Conversely, in stable environments, increasing the second time threshold can help reduce unnecessary notifications.

[0226]

[0193] Moreover, adapting the third time threshold based on environmental conditions can further optimize the method 100 for specific situations.

[0227]

[0194] In some embodiments, the method 100 may comprise a feedback mechanism that continuously monitors environmental conditions and adjusts the time thresholds accordingly. This can ensure that the method remains optimized for the current operating environment, leading to improved performance and reliability.

[0228]

[0195] According to an embodiment, the method 100 can comprise adapting the first distance threshold or the second distance threshold or the third distance threshold or the fourth distance threshold based on different types of environmental conditions.

[0229] 303774181.1

[0196] According to an embodiment, the method 100 involves determining the current environmental conditions and comparing them to previously stored environmental condition data. Based on this comparison, the system 200 can adjust the time or distance thresholds accordingly.

[0230]

[0197] For instance, in adverse weather conditions, such as heavy rain or fog, the first distance threshold maybe reduced to allow for greater safety in object detection and tracking.

[0231]

[0198] This feature enables the method 100 to maintain accurate safety action execution under various environmental conditions, thereby improving overall system 200 performance. Additionally, it allows for flexibility in adapting to changing environmental conditions, ensuring reliable operation in diverse environments.

[0232]

[0199] To ensure that the driver is aware of at least some risks of collisions, the system 200 can provide visual and audio and haptic warnings when a collision is deemed probable.

[0233]

[0200] According to an embodiment, the method 100 comprises providing a notification in an audible manner. Preferably, this audible notification can take the form of a sound or series of sounds that are distinct and easily recognizable. The method 100 may also comprise providing a visual notification. This visual notification can be presented on the screen 231, in the form of an icon, a message, a flash of the entire screen 231 in a bright color such as red, for example, or other graphical representation. Furthermore, according to another embodiment, the method 100 may involve providing a haptic notification. This haptic feedback can be delivered through various means, such as vibrations, pulses, or taps, and can be used to alert the user of an event or condition.

[0234]

[0201] Moreover, the method 100 may comprise determining which type of notification to use based on the nature of the event or condition being signaled. For instance, an audible alert may be more appropriate for urgent situations, while a visual notification may be more suitable for less time-sensitive events. In summary, the method 100 can comprise providing notifications in audible, visual, or haptic forms, with each type of notification offering distinct advantages for different use cases and environments. The choice of which form(s) to use may depend on the nature of the event or condition being signaled.

[0235] 303774181.1

[0202] According to an example, adjusting the watercraft's heading comprises controlling a waterjet thrust direction in different directions to control speed, acceleration, heading, and attitude for maneuvers.

[0236]

[0203] By controlling a watercraft's heading through adjusting the thrust direction of a waterjet, the present technology enables precise and agile maneuvers. This is particularly useful in applications where quick turns or changes in direction are required, such as in recreational activities or when the watercraft must avoid a collision.

[0237]

[0204] Adjusting the waterjet thrust direction enables optimized control of speed and acceleration, leading to increased control in watercraft propulsion. This can allow the watercraft to be stopped quickly and efficiently by changing the direction of the waterjet.

[0238]

[0205] According to an example, the present technology further comprises the application of force resistance feedback on the throttle lever when the motor power output is reduced below the predetermined limit.

[0239]

[0206] The application of force resistance feedback on the throttle lever provides an enhanced motor control system. By detecting when the motor power output falls below a predetermined limit, this feature can apply resistance to the throttle lever, preventing the operator from inadvertently increasing motor power demand, but not blocking the operator though, should the operator, for example, need more throttle to perform a short burst for e.g. docking, safety reasons, waves, etc.

[0240]

[0207] Force resistance feedback on the throttle lever significantly improves operator safety. In situations where the motor power output is reduced to avoid an object, for example, this feature prevents the operator from unintentionally increasing motor demand.

[0241]

[0208] According to an embodiment, the watercraft 10 can have a reverse gate 86, which allows for fine control of the vehicle's dynamics. The system 200 can be configured to control the reverse gate 86 precisely over its entire range 15, allowing for smooth and fine control of the watercraft’s speed, deceleration, direction, and attitude.

[0242]

[0209] According to an example, and as represented by figure 15, the method 100 further comprises:

[0243] 303774181.1 a) Determining a limitation axis of intent 12 configured to establish a reference frame for collision avoidance calculations, and determining a penetration angle 15 and / or a penetration distance 16 between the watercraft 10 and the detected object 20 relative to the limitation axis of intent 12; and b) Determining a required power limitation based on the penetration angle 15 and the penetration distance 16 between the watercraft 10 and the detected object 20 relative to the intention axis.

[0244]

[0210] By determining a limitation axis of intent 12 and calculating the penetration angle 15 and distance 16 between the watercraft 10 and the detected object 20 relative to this axis 12, the present technology enables more accurate collision avoidance calculations. This is useful in dynamic aquatic environments where objects maybe moving unpredictably.

[0245]

[0211] According to an embodiment, and as illustrated by figures 14, 15 and 16, a limitation axis of intent 12 is determined to establish a reference frame for collision avoidance calculations. This limitation axis 12 represents the intended direction of movement of the watercraft 10 and serves as a basis for calculating potential closest distances also called close calls with surrounding objects 20. Additionally, the penetration angle or penetration distance between the watercraft 10 and a detected object 20 are determined relative to the limitation axis of intent 12. The penetration angle refers to the angle between the direction of the watercraft's intended movement and the direction of approach to the detected object 20, while the penetration distance represents the distance between the two entities along this angle.

[0246]

[0212] According to an embodiment, the required power limitation is determined based on the calculated penetration angle and distance. This adaptive power management feature allows for optimized navigation.

[0247]

[0213] Therefore, by determining a limitation axis of intent 12, the method 100 provides a stable reference frame for collision avoidance calculations, improving accuracy and reliability. The calculation of penetration angle and distance between the watercraft 10 and the object 20 relative to the intention axis 12 enables collision avoidance maneuvers. The required power limitation based on the penetration angle and distance ensures optimal motor adjustments for reducing the probability of a collision.

[0248] 303774181.1

[0214] The present technology's ability to calculate penetration angle and distance, as well as determine a required power limitation, provides the watercraft 10 with enhanced situational awareness. This information can be used by the operator or autonomous systems to make informed decisions regarding navigation and collision avoidance in real-time.

[0249]

[0215] The present technology's advanced collision avoidance calculations and adaptive power management capabilities contribute to safer and more efficient watercraft operation. By reducing the probability of collisions, the watercraft 10 can navigate through complex aquatic environments with confidence.

[0250]

[0216] According to an example, the present technology further comprises the integration of steering angle sensor data to enable the driver to adjust the vehicle's trajectory and align a new axis with the intended direction for proactive collision avoidance. The integration of steering angle sensor data allows for real-time monitoring of the vehicle's steering direction. By analyzing this data, the system can predict potential collisions and suggest corrective actions to the driver.

[0251]

[0217] According to an embodiment, the method 100 involves receiving data from various sensors installed on the vehicle. This data comprises, but is not limited to, steering angle sensor data, for example. The method 100 can incorporate the integration of the received steering angle sensor data. This integration enables the driver to adjust the vehicle's trajectory in real-time and align the new axis with the intended direction for proactive collision avoidance. For example, the method 100 utilizes the processing module 230 that processes the sensor data and generates appropriate control signals to adjust the steering angle of the watercraft 10. This allows the watercraft 10 to respond quickly to changing water conditions or unexpected obstacles. The method 100 may also comprise the use of additional sensors such as lidar, radar, or cameras to provide a more comprehensive view of the environment surrounding the watercraft 10. This data can be integrated with the steering angle sensor data to improve the accuracy and effectiveness of the collision avoidance system. The integration of steering angle sensor data in real-time enables proactive collision avoidance, improving safety for the user.

[0252]

[0218] The new axis refers to the imaginary line that represents the safest path for the vehicle to travel. By incorporating steering angle sensor data, the system 200 can determine the intended direction of the vehicle and align the new axis accordingly. This ensures that the

[0253] 303774181.1 driver receives accurate guidance and feedback, enabling them to maintain a safe driving position even in complex traffic situations or adverse weather conditions.

[0254]

[0219] According to an embodiment, and as illustrated by figures 14 to 16, the method 100 comprises receiving the position and steering angle of the watercraft 10, as well as the position and accelerometer value of the detected object 20. Preferably, the steering angle value is input with a gain (K) to determine the angle of the axis of intent 12. The detected object 20 is virtually projected onto the axis of intent 12 to obtain a longitudinal factor of limitation. The angular difference between the axis of intent 12 and the axis 22 between the watercraft 10 and the object 20 is used to derive an angular limitation. The longitudinal factor of limitation and the angular limitation are then multiplied together to yield a throttle limitation, see for example figure 16.

[0255]

[0220] Indeed, according to an embodiment, and as illustrated by figures 14, 15 and 16, the method 100 further comprises: a) receiving 310 the watercraft position, steering angle, accelerometer value and the detected object position; b) inputting 320 a steering angle value with a gain (K) to determine the angle 15 of the axis of intent 12; c) using 330 a projection of the detected object 20 on the axis of intent 12 to find a longitudinal factor of limitation; d) using 340 the angle difference between the axis of intent and the axis between the watercraft 10 and the detected object 20 to get an angular limitation; e) multiplying 350 the longitudinal factor of limitation with the angular limitation together to get a throttle limitation.

[0256]

[0221] By receiving and processing the watercraft position, steering angle, accelerometer value, and object position, the present technology enables precise determination of a triangular zone t figure 15 and calculation of angular and longitudinal limitations. According to an embodiment, the present technology is configured to calculate the axis of the detected object 20, the distance projection and the relative velocity.

[0257] 303774181.1

[0222] The use of both angular and longitudinal limitation factors allows the system 200 to account for potential collisions with objects in the watercraft's path. By calculating these limitations based on the object position and the angle difference between the axis of intent and the axis between the watercraft and the object, the method can provide timely throttle adjustments to prevent or mitigate collisions, enhancing overall safety for passengers and other watercraft users.

[0258]

[0223] The application of a gain (K) when inputting a steering angle value allows the system to adapt to different operating conditions and environments. This flexibility ensures that the watercraft 10 responds appropriately to various situations providing a safer, more comfortable, and efficient user experience.

[0259]

[0224] The present technology's ability to quickly process and utilize the received data enables real-time adjustments to the watercraft's steering and throttle. This responsiveness is useful for collision avoidance and maintaining optimal control in dynamic environments, such as busy waterways or open waters with changing weather conditions.

[0260]

[0225] According to an embodiment, a limitation of throttle / speed / torque can be applied to the watercraft for any variety of reasons, such as collision avoidance. The technology can be configured to provide haptic feedback to the driver through the throttle lever when a limitation is detected. Two haptic feedback modes can be integrated into the lever: vibration and resistance.

[0261]

[0226] According to an embodiment, the present technology is applicable to a watercraft 10 that has a limitation to its throttle, speed, power, torque, RPM, acceleration, etc., which we will refer to as "throttle" for simplicity. According to an embodiment, the present technology includes at least four methods of overriding a throttle limitation: a) The driver may use the brake system to override a throttle limitation, particularly in watercraft that require throttle to brake faster. b) During the transition from a throttle limitation to a zero throttle limitation scenario, a rising time can be applied to the regain of throttle to ensure a smooth transition between phases.

[0262] 303774181.1 c) A "battery reserve of power” is proposed, which allows the driver to override the throttle limitation by using an amount of power x time. For example, when limited to 15% throttle, the driver begins with a reserve of lOOHp x 3 seconds and can use it in various ways (e.g., 300hp x 1 second or 50hp x 6 seconds). Once depleted, a recharge timer starts filling up the reserve for reuse and is hereinafter referred to as battery reserve of power. d) The last method applies a negative gain to the throttle limitation based on the steering angle of the driver, with zero gain at a steering angle of 0 and maximum gain at maximum / minimum steering angles (left / right).

[0263]

[0227] According to an embodiment, the technology ensures that power restoration to the driver through the override is not abruptly set to 100%. It intelligently manages the throttle / speed / torque restoration process based on the driver's input and previous limitations. The driver can quickly regain power through the activation of the vehicle's intelligent brake and reverse system, which enhances the driver's ability to maneuver and respond effectively.

[0264]

[0228] According to an example, the present technology comprises offering throttle limitation override proceedings, these throttle limitation override proceedings comprising: a) Using the reverse gate 86 to override a limitation, during the transition from a throttle limitation to a zero throttle limitation scenario; or b) Using a "battery reserve of power” configured to be used to override the throttle limitation for a predetermined limited time.

[0265]

[0229] This feature allows for the override of a throttle limitation during the transition to a zero throttle scenario by employing a braking system associated with a reverse gate 86 of the watercraft 10. By applying a reverse thrust with the reverse gate 86, the present technology allows to reduce the speed or even stop the watercraft without stopping the propulsion of water.

[0266]

[0230] The "battery reserve of power” provides drivers with a controlled amount of power that can be used to override limitations to temporarily override the limitation. The reserve is composed of an amount of power allowing the driver to use it in various ways. Once depleted,

[0267] 303774181.1 a recharge timer refills the reserve, enabling its reuse. This feature allows drivers control while also enforcing rules and regulations.

[0268]

[0231] According to an embodiment, during docking proceedings, as illustrated by figure 12, the present technology can help maintain a safe distance from a trailer 14, the dock and / or other boats, with the calculated distance between the detected object 20 and the watercraft 10 displayed on the dashboard or heads-up display.

[0269]

[0232] According to an embodiment, the method 100 is implemented at least partially by the system 200.

[0270]

[0233] According to an embodiment, the system 200 comprises the perception-based sensor 210 mounted on the watercraft 10 and configured to acquire data from its surroundings. The system 200 also comprises the neural network module 220 with the pre-trained neural network, which is responsible for detecting an object 20 within a predetermined area located in front of the watercraft 10 using the acquired data from the perception-sensor 210. The system 200 further comprises the processing module 230 configured to perform several functions.

[0271]

[0234] According to an embodiment, the processing module 230 is configured to acquire data from the perception-based sensor 210 and to send it to the neural network module 220 for object detection.

[0272]

[0235] According to an embodiment, in response to the detection of an object 20 by the neural network module 220, the processing module 230 is configured to extract a first set of information regarding the detected object 20 using this data. This first set of information includes an estimation of the detected object's attitude, such as pitch, roll, or yaw.

[0273]

[0236] According to an embodiment, the processing module 230 is configured to acquire a second set of information about the watercraft 10 from a set of sensors of the watercraft 10. The processing module 230 is also configured to compute a third set of information using both the first and second sets of data. This third set of information includes the predicted trajectory of the detected object 20, the predicted trajectory of the watercraft 10, and the relative speed between them.

[0274] 303774181.1

[0237] Based on this computed third set of information, the processing module 230 is configured to determine if a close call distance (CCD) exists between the detected object 20 and the watercraft 10, i.e. to determine the closest distance.

[0275]

[0238] In response to the existence of a potential CCD point, the processing module 230 is configured to execute at least one specific safety action based on the computed third set of information.

[0276]

[0239] According to an embodiment, the system 200 can be configured to cooperate with or control multiple watercraft. The system 200 can, for example, be embodied in a configuration where it controls multiple watercraft to maintain safe distances from each other and reduce collisions. This embodiment enables the system 200 to coordinate the movements of multiple watercraft, adjusting their headings, speeds, and attitudes as needed to ensure safe separation.

[0277]

[0240] In this configuration, the system 200 can, for example, receive data from perception sensors on each watercraft, determining the relative positions and trajectories of all participating vessels as well as objects. The system 200 then can be configured to calculate optimal control inputs for each watercraft to maintain safe distances and reduce the probability of collisions between them and between objects that are not controlled by the system 200, taking into account factors such as speed, direction, and maneuverability.

[0278]

[0241] According to an embodiment, the present technology is designed to be compatible with a wide range of watercraft, including those without advanced communication or sensor systems. It does not require any communication with other vessels or objects to function.

[0279]

[0242] Modifications and improvements to the above-described implementations of the present technology may become apparent to those skilled in the art. The foregoing description is intended to be exemplary rather than limiting. The scope of the present technology is, therefore, intended to be limited solely by the scope of the appended claims.

[0280] 303774181.1

Claims

What is claimed is:

1. A watercraft comprising: a hull; a deck mounted on the hull; a motor connected to at least one of the deck and the hull; a propulsion system operatively connected to the motor; a steering handlebar operatively connected to the deck; a front splash guard mounted to the deck forward of the steering handlebar; a perception-based sensor mounted: behind the front splash guard; in front of or on the steering handlebar; and above a waterline of the watercraft.

2. The watercraft according to claim 1 wherein the perception-based sensor is located on a front cowling of the watercraft.

3. The watercraft according to claim 1, further comprising a housing housing the perception-based sensor, the housing being configured to protect the perception-based sensor from water.

4. The watercraft according to claim 1, wherein the perception-based sensor is an optical sensor.

5. The watercraft according to claim 1, wherein the perception-based sensor is a camera capable of capturing at least one of images or videos in real time.

6. The watercraft according to claim 1, further comprising a processing module connected to the perception-based sensor, the processing module being responsible for analyzing data from the perception-based sensor and for determining the presence and location of objects in a vicinity of the watercraft.303774181.

17. The watercraft according to claim 1, comprising a system for perception-based trajectory determination connected to the perception-based sensor.

8. The watercraft according to claim 7 wherein the system for perception-based trajectory determination for a watercraft comprises: a neural network module comprising a pre-trained neural network and being configured to detect an object within a predetermined area, the predetermined area being located around the watercraft, using the acquired data; a processing module configured to: acquire data from the perception-based sensor; send the acquired data to the neural network module; detect an object located in the predetermined area of observation using acquired data from the perception-based sensor; extract a first set of information regarding the detected object, the first set of information being configured to be used to determine a predicted trajectory of the detected object, the first set of information comprising: a relative distance between the detected object and the watercraft; a detected object position based on the coordinates of the detected object within a field of view of the perception-based sensor; and a relative bearing between the detected object and the watercraft; acquire a second set of information about the watercraft from a set of sensors of the watercraft; compute a third set of information using the first set of information and the second set of information secondary information, the third set of information comprising a predicted trajectory of the detected object and a predicted trajectory of the watercraft and a relative speed between the detected object and the watercraft; determine a closest distance between the detected object and the watercraft exist, based on the predicted trajectory of the detected object, the predictive trajectory of the watercraft and the relative speed between the detected object and the watercraft; and in response to the determination of the closest distance, execute a safety action among a set of safety actions according to a predetermined set of safety rules.303774181.

19. The watercraft according to claim 8, wherein the processing module is further configured to provide a user of the watercraft with real-time information about the detected object.

10. The watercraft according to claim 8, wherein the safety action is taken from a set of safety actions according to a predetermined set of safety rules.

11. The watercraft according to claim 8, wherein, in response to the determination of the closest distance, the processing module is configured to determine at least one of: a direct distance between the actual position of the watercraft and a point of the predicted trajectory of the watercraft corresponding to the point of the closest distance; or a time to the closest distance corresponding to the time to reach a point of the predicted trajectory of the watercraft corresponding to the point of the closest distance.

12. The watercraft according to claim 11, wherein the processing module is further configured to: warn the user via a notification in response to at least one of the time to the closest distance being smaller than a predetermined first time threshold or the direct distance being smaller than a first predetermined distance threshold; reduce motor power output below a predetermined limit in response to at least one of the time to the closest distance being smaller than a predetermined second time threshold or the direct distance being smaller than a second predetermined threshold; and automatically brake the watercraft using a reverse gate in response to at least one of the time to the closest distance being smaller than a predetermined third time threshold or the direct distance being smaller than a third predetermined threshold.

13. The watercraft according to claim 1 wherein the perception-based sensor comprises an additional camera, and wherein the camera comprises a first field of view, and the additional camera comprises a second field of view, the first field of view being wider than the second field of view.

14. The watercraft according to claim 1 wherein the perception-based sensor is mounted to form an angle Al with the waterline, the angle Al being comprised between -25 degrees and +25 degrees.303774181.

115. The watercraft according to claim 1 wherein, the first set of information comprises an estimation of the detected object attitude comprising an estimation of at least one of: a pitch of the detected object; a roll of the detected object; or a yaw of the detected object.

16. The watercraft according to claim 1, comprising a water jet propulsion system for generating a thrust, and a reverse gate configured to redirect the thrust toward a front of the watercraft.303774181.1

Images