Self-driving vehicle, system and method for loading and unloading

Abstract

The invention relates to a method for loading and unloading a truck by means of a self-driving vehicle (SDV) comprising the steps of: (a) navigating a self-driving vehicle (SDV) up to a truck; (b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of a cargo space; (c) navigating the self-driving vehicle into the cargo space, until the self-driving vehicle has reached a suitable location for unloading or loading a load; (d) unloading or loading the load; and (e) navigating the self-driving vehicle out of the cargo space; wherein the navigating and / or maneuvering is performed by means of a localization camera and a plurality of safety sensors, said safety sensors comprising three or more laser sensors and two or more three-dimensional (3D) cameras.

Description

[0001] SELF-DRIVING VEHICLE, SYSTEM AND METHOD FOR LOADING AND UNLOADING

[0002] TECHNICAL FIELD

[0003] The invention relates to a self-driving vehicle, as well as systems and methods for controlling self-driving vehicles.

[0004] PRIOR ART

[0005] Self-driving vehicles (SDVs) are known in many variations from the prior art. SDVs are already commonly used in industrial and logistics environments, but these environments inherently present a number of challenges that hamper the smooth operation of SDVs.

[0006] An SDV typically uses a variety of technologies that attempt to enable tasks to be performed without human intervention. A widely used technology for navigation, for example, is the use of lines or magnetic strips on the floor, which the SDV follows, as described in WO 2017 / 044522. More advanced systems use cameras to create a detailed map of the environment, as described in, for example, EP 2987761. Although these prior art systems operate acceptably in static environments, they struggle with dynamic environments where people and unexpected obstacles may be present. Moreover, environments with poor visibility or limited space also pose difficulties for SDVs. In particular, the loading and unloading of trucks poses the necessary challenges in this context.

[0007] Cargo holds of trucks are usually very dark, and are limited in space, i.e., loads need to be positioned very precisely and close together. Traditional visual systems, such as cameras, often underperform as a result of poor lighting conditions. Consequently, the localization and navigation of the SDV inside the truck may be inaccurate, potentially leading to inefficient loading and possible damage to the cargo or the truck.

[0008] Another problem is the detection and handling of obstacles within the cargo space. Obstacles such as improperly placed pallets, tools, other objects, or people can disrupt operations and cause hazardous situations. Existing systems generally struggle to respond quickly and accurately to such obstacles, which can lead to emergency stops and delays, or even accidents, during the loading and unloading process.

[0009] The present invention aims to solve at least some of the above problems or drawbacks.

[0010] SUMMARY OF THE INVENTION

[0011] The current invention relates to a method for loading a truck using a self-driving vehicle (SDV) according to claim 1.

[0012] The method has the advantage that the self-driving vehicle (SDV) can navigate and / or maneuver efficiently and under control both in and out of the cargo space of the truck. Loading a truck using this method is therefore exceptionally efficient and precise. The combination of sensors as described herein further allows obstacles located in the vicinity of the self-driving vehicle to be quickly and efficiently detected.

[0013] Preferred embodiments of the method are set forth in dependent claims 2 to 9.

[0014] A further aspect of the present invention relates to a method for unloading a truck using a self-driving vehicle (SDV) according to claim 10. The unloading of a truck using this method is consequently exceptionally efficient and accurate.

[0015] Preferred embodiments of the method are set forth in dependent claims 11 to 13.

[0016] Further aspects of the invention relate to a self-driving vehicle (SDV) according to claim 14, and a system for controlling a self-driving vehicle according to claim 15.

[0017] DESCRIPTION OF THE FIGURES

[0018] Figure 1 shows a perspective view of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention.

[0019] Figure 2 shows another perspective view of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention. Figure 3 shows a front view of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention.

[0020] Figure 4 shows a rear view of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention.

[0021] Figure 5 shows a side view of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention.

[0022] Figure 6 shows a top view of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention.

[0023] DETAILED DESCRIPTION

[0024] In a first aspect, the present invention relates to a self-driving vehicle (SDV) comprising: a vehicle base, a load carrier, which load carrier is coupled to said vehicle base, an industrial computer unit, a programmable logic controller (PLC), a localization camera and a plurality of safety sensors, which industrial computer unit, programmable logic controller (PLC), localization camera and sensors are coupled to the vehicle base, wherein the industrial computer unit is electronically coupled to the programmable logic controller (PLC), and wherein said localization camera and said safety sensors are electronically coupled to the industrial computer unit and / or the programmable logic controller (PLC), and wherein said safety sensors comprise two or more laser sensors, two or more ultrasonic sensors, and two or more three-dimensional (3D) cameras.

[0025] Unless otherwise defined, all terms used in the description of the invention, including technical and scientific terms, have the meanings as commonly understood by a person skilled in the art to which the invention pertains. For a better assessment of the description of the invention, the following terms are explicitly explained.

[0026] "A," "an," and "the" as used herein refers to both singular and plural referents unless the context clearly dictates otherwise. For example, "a segment" means one or more than one segment. The terms "comprise," "comprising," "consist of," "consisting of," "provided with," "include," "including," "contain," "containing," are synonyms and are inclusive or open terms that indicate the presence of what follows, and which do not exclude or prevent the presence of other components, characteristics, elements, members, steps, as known from or disclosed in the prior art.

[0027] Quoting numeric intervals by the endpoints includes all integers, fractions, and / or real numbers between the endpoints, including those endpoints.

[0028] The self-driving vehicle (SDV) according to the present invention has the advantage that the combination of sensors as described herein allows obstacles located in the vicinity of the self-driving vehicle to be detected quickly and efficiently. The combination of different types of sensors provides redundancy and robustness in obstacle detection. This allows the vehicle to operate more safely in dynamic environments such as warehouses and production facilities where the presence of people and unexpected obstacles poses a risk. Following the detection of obstacles, the self-driving vehicle can potentially initiate certain control interventions, such as deviating from a planned navigation path, decelerating, accelerating, and / or coming to a stop. This rapid and efficient obstacle detection can thus not only lead to more efficient driving behavior, but also enhances the safety of persons in the vicinity.

[0029] In the context of the present invention, the term "vehicle base" refers to the fundamental frame or chassis of the self-driving vehicle on which all other components are mounted.

[0030] The term "load carrier" refers to a structure or platform specifically designed to carry and transport loads such as pallets.

[0031] The term "industrial computer unit" denotes a computer that is capable of performing complex calculations and that constitutes the central processing and control unit for the self-driving vehicle.

[0032] A "programmable logic controller (PLC)" is to be understood as an electronic device that is designed for industrial automation and that is responsible for controlling machines and processes by executing specific logical operations and commands. In the context of the invention, the term "localization camera" is to be interpreted as a camera that is used to accurately determine the position of the self-driving vehicle in its environment by means of visual data.

[0033] The terminology "safety sensor" refers to a group of sensors intended for detecting obstacles and persons in the immediate vicinity of the self-driving vehicle to prevent collisions and ensure safety. In particular, in the context of the invention, these safety sensors are laser sensors, ultrasonic sensors, and three-dimensional (3D) cameras.

[0034] A "laser sensor" is to be understood as a sensor that uses laser beams to measure the distance to objects and thus detect obstacles in a specific plane.

[0035] The term "ultrasonic sensor" denotes a sensor that uses high-frequency sound waves to measure the distance to objects and detect obstacles, particularly useful for detecting objects at close range.

[0036] A "three-dimensional (3D) camera," in the context of the invention, should be understood as a camera that is capable of obtaining depth information by creating a three-dimensional image of the environment, allowing the vehicle to detect not only the presence but also the dimensions and shape of objects.

[0037] Preferably, said laser sensors are coupled to the front side and / or to the lateral sides of the vehicle, said ultrasonic sensors are coupled to the rear side of the vehicle, and said three-dimensional (3D) cameras are coupled to the front side of the vehicle.

[0038] In the context of the present invention, the terms "front side," "rear side," and "lateral sides" are used according to the direction of travel of the self-driving vehicle. Thus, the load carrier is located at the rear side of the self-driving vehicle.

[0039] Due to the strategic placement of the laser sensors on the front side and / or lateral sides, ultrasonic sensors on the rear side, and three-dimensional (3D) cameras on the front side, the vehicle can accurately detect surrounding obstacles. This leads to improved situational awareness and increased safety.

[0040] According to a further or alternative embodiment, said safety sensors are configured to monitor a monitoring field that is at least 10% larger than a safety field around the self-driving vehicle. In the context of the present invention, several terms are used that designate a field around the self-driving vehicle. The term "field of view" refers to the maximum perceivable zone of the various safety sensors. The field of view is thus generally limited by the technical characteristics of the various sensors and / or the manner in which they are mounted on the self-driving vehicle. The so-called "safety field" of the sensors is a specific part of the field of view used for safety purposes. This is thus the zone in which safety sensor monitoring is required and / or mandatory for the purpose of preventing hazardous situations. Typically, the required and / or mandatory dimensions of this safety field are defined in the applicable safety standards and guidelines. When an object or person enters this safety field, the system will undertake a safety action, such as slowing down or stopping the vehicle, to prevent an accident. The safety field is therefore a subset of the field of view and is designed to comply with safety standards and guidelines. A field that, in terms of size, is situated between the safety field and the field of view, is the so-called "monitoring field." In the context of the present invention, the monitoring field is the effective zone around the self-driving vehicle that is monitored by the safety sensors.

[0041] The self-driving vehicle as described herein has the advantage that the various sensors of the self-driving vehicle not only monitor the direct safety field, but also scan a larger area within the field of view, namely the monitoring field. This enables the system to detect potential obstacles and hazards at an early stage and thus proactively adjust the control of the vehicle, which can reduce the need for sudden emergency stops.

[0042] Preferably, said safety sensors are configured to monitor a monitoring field that is at least 15% larger than a safety field around the self-driving vehicle, more preferably, said safety sensors are configured to monitor a monitoring field that is at least 20% larger than the safety field around the self-driving vehicle, even more preferably, said safety sensors are configured to monitor a monitoring field that is at least 25% larger than the safety field.

[0043] Proactively adjusting the control of the self-driving vehicle based on monitoring of the monitoring field, comprises, according to some embodiments, deviating from a planned navigation path, decelerating, accelerating, and / or bringing the self-driving vehicle to a stop. According to a further or alternative embodiment, said laser scanners together have a two-dimensional (2D) field of view, which field of view extends in a plane substantially parallel to a ground surface, preferably at a distance comprised between 5 and 20 cm from said ground surface.

[0044] The laser scanners that provide a two-dimensional field of view at a height of 5 to 20 cm from the ground surface make it possible to effectively detect obstacles at a substantial height above the ground surface. This is particularly advantageous for preventing collisions with objects that might be overlooked by other sensors.

[0045] Preferably, said laser scanners together have a two-dimensional (2D) field of view, which field of view extends at a distance comprised between 8 and 18 cm from said ground surface, more preferably comprised between 10 and 15 cm, most preferably comprised between 12 and 14 cm from said ground surface.

[0046] According to some embodiments, said laser scanners together have a two- dimensional (2D) field of view, which field of view extends in a plane substantially parallel to the ground surface, which field of view extends in this plane over a maximum distance of 15 meters around the self-driving vehicle, preferably a maximum distance of 12 meters, more preferably a maximum distance of 10 meters around the self-driving vehicle.

[0047] According to a further or alternative embodiment, the vehicle base of the self-driving vehicle (SDV) is provided with one or more cut-outs at the location of said laser scanners. These cut-outs allow the laser scanners to perceive a wider two- dimensional (2D) field of view, where they would otherwise be obstructed by the vehicle base.

[0048] According to some embodiments, said laser scanners together have a two- dimensional (2D) field of view that extends over an angle comprised between 220° and 280° around the self-driving vehicle. More preferably, said laser scanners together have a two-dimensional (2D) field of view that extends over an angle comprised between 230° and 280° around the self-driving vehicle. Still more preferably, said laser scanners together have a two-dimensional (2D) field of view that extends over an angle comprised between 240° and 280° around the self-driving vehicle, further preferably over an angle comprised between 250° and 280°, and most preferably over an angle comprised between 260° and 280°. The self-driving vehicle as described herein has, thanks to the extended field of view of the sensors, improved situational awareness, which leads to higher safety.

[0049] According to a further or alternative embodiment, the load carrier comprises one or more forks, and the ultrasonic sensors are located in these one or more forks, preferably fork tips, of the load carrier. Preferably, an ultrasonic sensor is located in each fork tip of the load carrier.

[0050] According to a further or alternative embodiment, the ultrasonic sensors have a detection range in the form of a three-dimensional (3D) sound beam, that points rearward in the zone behind the fork tips of the load carrier.

[0051] The ultrasonic sensors as described herein ensure that the sensors are specifically optimized for critical reverse maneuvers, thereby improving the safety and precision during these operations. This has the advantage that the ultrasonic sensors provide accurate and reliable detection during backward movements in specific maneuvering zones. This increases the safety and efficiency when performing these maneuvers. Preferably, the ultrasonic sensors are only active when the self-driving vehicle is moving backward, which is only permitted in special maneuvering zones, such as for picking up and setting down pallets, parking, or charging.

[0052] According to a further or alternative embodiment, the self-driving vehicle (SDV) further comprises a pallet camera. This pallet camera is capable of detecting a pallet when one is present behind the self-driving vehicle (SDV). This enables the selfdriving vehicle to efficiently load and / or unload pallets.

[0053] Preferably, said pallet camera is a LiDAR camera. A "LiDAR camera" generates depth images and is used in the present context for the accurate detection of pallets. In an alternative preferred embodiment, said pallet camera is a three-dimensional (3D) camera.

[0054] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises one or more safety encoders and a safety PLC. The safety encoders are capable of measuring the speed and steering angle of the self-driving vehicle (SDV). According to some embodiments, the measured values from the safety encoders are processed by the safety PLC, which safety PLC defines a safety field and / or monitoring field within the field of view of the safety sensors. This concerns a dynamic adaptation of the sensors to the driving conditions, which improves the safety and precision of obstacle detection.

[0055] The advantages of this embodiment are that the continuous monitoring of speed and steering angle by safety encoders and the dynamic adaptation of the scanner and safety fields by the safety PLC ensure optimal and reliable obstacle detection. This leads to improved safety and efficiency in the operation of the SDV, because the sensors are always optimally configured for the current driving conditions.

[0056] Preferably, said safety PLC is a safety PLC with performance level PL-d.

[0057] In the context of the present invention, the term "performance level PL-d" refers to a classification within the ISO 13849-1 standard, which determines the safety of control systems for machinery. This standard describes the requirements for the design and integration of safety systems and provides a way to evaluate and verify the reliability and performance of safety functions.

[0058] According to a further or alternative embodiment, said laser scanners are configured for identifying a type of pallet, preferably by monitoring one or more load detection fields.

[0059] A "load detection field" is a subset of the field of view of the laser scanners, and is designed to detect a pallet located on the load carrier of the self-driving vehicle and to identify the type of pallet.

[0060] Preferably, the one or more load detection fields are substantially directed towards the rear of the vehicle, more preferably towards a load on the load carrier, more preferably towards a pallet on the load carrier.

[0061] According to some embodiments, the type of pallet comprises a Euro-pallet or an industrial pallet. Euro-pallets have standard dimensions of 800 x 1000 mm, and industrial pallets have standard dimensions of 1000 x 1200 mm.

[0062] According to some embodiments, the one or more load detection fields are designed for detecting and identifying Euro-pallets, preferably within the width of the vehicle base, and / or for detecting and identifying industrial pallets, preferably with a margin of 10 cm.

[0063] According to some embodiments, each of said laser scanners is configured to monitor a first load detection field that is designed for detecting and identifying Euro-pallets, preferably within the width of the vehicle base, and to monitor a second load detection field that is designed for detecting and identifying industrial pallets, preferably with a margin of 10 cm. Each of said laser scanners is thus individually capable of detecting whether the present pallet is a Euro-pallet or an industrial pallet. Thus, the laser scanners are able to verify with respect to each other whether a correct type of pallet was identified.

[0064] Preferably, said laser scanners switch repeatedly between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill. More preferably, said laser scanners repeatedly switch between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill at a frequency comprised between 1 and 2 Hz, more preferably between 1.2 and 2 Hz, between 1.4 and 2 Hz, between 1.6 and 2 Hz, or between 1.8 and 2 Hz. Most preferably, said laser scanners repeatedly switch between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill at a frequency of 2 Hz. This rapid switching allows load detection to take place efficiently just before lifting a pallet and / or during the lifting of a pallet.

[0065] In the context of the present invention, the term "standstill" designates a state of the self-driving vehicle, wherein the vehicle is moving at a speed of less than 0.1 m / s.

[0066] According to some embodiments, said laser scanners switch at least 5 times between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill, preferably at least 10 times, more preferably at least 15 times, or most preferably at least 20 times.

[0067] According to a further or alternative embodiment, the safety PLC is configured to adapt the safety fields based on a type of pallet identified by said laser scanners. In particular, when an industrial pallet is identified, the safety fields are enlarged to compensate for the larger dimensions of the pallet. This prevents hazardous situations where incorrect safety fields are used. According to some embodiments, the identification of a pallet type by said laser scanners meets performance level PL-d when the self-driving vehicle is at a standstill, and performance level PL-b when the self-driving vehicle is in motion.

[0068] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises at least one laser scanner, preferably a LiDAR sensor, placed in a fork tip of the load carrier. Preferably, a laser scanner, preferably a LiDAR sensor, is placed in each fork tip of the load carrier, such that the vehicle obtains detailed monitoring of the space behind the forks during backward movement. Preferably, no ultrasonic sensors are present in the fork tips in this embodiment.

[0069] Thus, an independent aspect of the invention also concerns a self-driving vehicle (SDV) comprising: a vehicle base, a load carrier, which load carrier is coupled to said vehicle base, an industrial computer unit, a programmable logic controller (PLC), a localization camera and a plurality of safety sensors, which industrial computer unit, programmable logic controller (PLC), localization camera and sensors are coupled to the vehicle base, wherein the industrial computer unit is electronically coupled to the programmable logic controller (PLC), and wherein said localization camera and said safety sensors are electronically coupled to the industrial computer unit and / or the programmable logic controller (PLC), and wherein said safety sensors comprise three or more laser sensors and two or more three-dimensional (3D) cameras.

[0070] The LiDAR sensors are preferably implemented as planar LiDAR sensors, and more particularly as sensors which are configured for safety applications with at least Performance Level PL-c, more preferably Performance Level PL-d. The LiDAR sensors are preferably configured with a response time of at most 228+6 ms, more preferably at most 114+6 ms.

[0071] According to some embodiments, the detection field of each LiDAR sensor is configurable into sixteen detection areas, wherein each detection area comprises two zones. These zones are positioned such that they fully or partially cover the width of the load carrier, wherein the detection fields of the left and right LiDAR sensors overlap to ensure redundant safety. In particular, it is provided that when reversing with a load, such as a longside pallet of 1200 mm width, the detection fields overlap across the entire width of the vehicle and the load.

[0072] According to some embodiments, a configuration is provided with multiple types of fields, including fields for:

[0073] 500 mm protection (longside / IND / EUR) left and right;

[0074] 500 mm confined long fields (EUR / IND / LONG) left and right;

[0075] 180 mm confined fields for end position just before wall contact; High-speed detection fields between 1300 mm and 2000 mm; Pallet detection fields for EUR, IND, and longside applications.

[0076] This field structure allows the SDV to adequately select safety fields in different zones and at different speeds, for example, when driving in a truck or performing loading or unloading operations.

[0077] The detection fields of the LiDAR sensors preferably overlap in the middle, such that the entire width of the load carrier is monitored, even when a load is present on the forks. In particular, when a pallet of 1200 mm width is being transported, the overlapping detection fields ensure that the area behind the vehicle is fully monitored, even when reversing at nominal speed, for example, above 0.3 m / s.

[0078] The LiDAR sensors are used in particular for space monitoring when reversing in narrow areas, such as truck cargo spaces. In this context, the strategic placement of the sensors allows for early detection of persons or obstacles, even in situations where other sensors, such as visual cameras, underperform due to low light conditions. By temporarily "muting" the detection fields of the LiDAR sensors when approaching fixed walls, unnecessary stops are avoided, while still complying with safety regulations.

[0079] According to some embodiments, the LiDAR sensors are also used to detect the pallet type while driving into a pallet. Here, a detection strategy is applied whereby three characteristic detection events are observed during driving in. These are generated by the tripping of three consecutive zones per fork, which corresponds to the presence of the three rows of blocks of a typical pallet. This procedure is performed on both sides, and when a threefold trip structure is established on both sides within a defined time interval and travel distance, the fork load can be identified as correct. According to a further or alternative embodiment, the self-driving vehicle (SDV) is configured to verify the type of pallet present on the forks, particularly when the SDV is already loaded at startup or when a pallet has been manually placed on the forks. In such cases, wherein the SDV does not have historical field data from driving into the pallet, a load type verification procedure is activated by the safety PLC. This procedure comprises activating specific verification fields, preferably provided in the front and lateral safety sensors, which fields are configured as detection fields with so-called "additional wings" for distinguishing longside pallets.

[0080] Preferably, this verification is performed after the pallet has been lifted to a predefined height, determined via a lift position sensor with at least performance level PL-d. If the width profile of the load on the forks corresponds to the profile of a longside pallet, this is validated by the tripping of one or more specific control fields, designated as validation fields. Based on this validation, the system automatically switches to the correct safety field configuration (e.g., EUR / IND versus LONG) for further travel or loading / unloading operations.

[0081] If a mismatch is detected between the expected and the actually detected carrier type, the SDV manager will block the execution of certain behaviors, such as driving, performing rack drops, or initiating loading or unloading procedures, until a correct identification of the carrier type has been confirmed or manual confirmation has been received via the user interface.

[0082] According to some embodiments, when the SDV is in service, these detection fields are permanently activated to ensure continuous load monitoring. If a pallet is already present on the forks at startup, the SDV can optionally allow a load type to be selected via the user interface, which is subsequently checked by a separate validation field.

[0083] According to a further or alternative embodiment, the load type identification (e.g., EUR versus IND) is particularly relevant for automatically adapting the active safety fields. If, for example, an industrial pallet is detected, a wider field is selected than for a Euro-pallet. Preferably, a so-called W2 safety zone is also monitored at the rear of the vehicle to determine whether or not the load is positioned against the chassis of the vehicle. According to a further or alternative embodiment, the self-driving vehicle (SDV) is configured to automatically activate the correct safety fields during backward movement based on:

[0084] - the type and width of the load;

[0085] - the detected presence of obstacles via one or more safety sensors;

[0086] - the driving speed and steering angle (for example, limited to ±5° in narrow aisles);

[0087] - the distance to fixed objects such as a wall or an already loaded pallet.

[0088] For example, if a Euro-pallet or no load is present on the forks, a standard field of 0.8 m width is selected. For an industrial pallet, this becomes 1.0 m, and for a longside pallet, 1.2 m. As soon as an obstacle or wall is detected at a distance of 500 mm, a switch is made to a 180 mm field that is exclusively monitored by the LiDAR sensors in the fork tips. When the tripping of the 180 mm field is detected, the LiDARs can be deactivated (muted) to allow for collision-free positioning without exceeding ISO safety standards. In this field, it is assumed that no person can be present, such that the LiDAR sensors can be temporarily deactivated. This makes it possible to position the load up against a wall or another pallet without compromising safety.

[0089] According to some embodiments, in narrow zones, as defined in the safety configuration of the system, the lateral fields are shifted inward by a maximum of 18 cm ("narrow mode") in order to maintain maneuverability. According to a further or alternative embodiment, the self-driving vehicle (SDV) is configured to selectively activate different frontal safety fields based on the current driving speed and lift height of the load. In particular, at low speeds, preferably below 0.4 m / s, the foot margin at the front of the vehicle is deactivated, and only the margins of the frontal safety sensors are taken into account. At higher speeds, preferably above 0.4 m / s, an additional foot margin is activated for foot detection near the ground surface around the vehicle, depending on the type and dimensions of the pallet on the forks. In particular, when driving with a Euro-pallet, a wider foot margin is provided by default. If the pallet is lifted higher than 90 mm, this additional foot margin is canceled for industrial or longside pallets, in accordance with the application of projection methods as defined in the foot margin configuration.

[0090] In the context of the present invention, the term "foot margin" denotes a safety zone located at the front of the self-driving vehicle (SDV), at the level of the ground surface. The foot margin aims to detect people who are right in front of the vehicle, particularly in the area where a foot might be located. The dimensions and activation of the foot margin depend on operational parameters such as speed, lift height of the load, and the type of load. Preferably, the foot margin is deactivated or minimized at low speeds, while at higher speeds, and especially at lower lift heights, a larger foot margin is applied to avoid collisions with people. The projection of the foot margin preferably occurs through algorithmic extension of the detection field of the frontal safety sensors.

[0091] The selection and activation of these frontal safety fields is preferably done dynamically and in real time based on input from a lift position sensor (preferably with performance level PL-d) and the current speed measurement, and is managed via the SDV manager or safety PLC.

[0092] The various embodiments of the self-driving vehicle as described herein preferably include a drive system.

[0093] According to a further or alternative embodiment, the self-driving vehicle further comprises one or more force sensors. These force sensors enable the self-driving vehicle to detect physical interactions with it, and to potentially respond thereto.

[0094] This is referred to as "force feedback," i.e., a mechanical or electronic system that feeds back forces or resistive forces experienced by a vehicle, such as a self-driving vehicle (SDV), to the control systems of the vehicle. This enables the vehicle to react in real-time to physical interactions with its environment by making adjustments to the movement or navigation parameters.

[0095] Force feedback can, according to some embodiments, also be measured without separate force sensors. According to some embodiments, force feedback is measured as an increase in load on the drive system, for example, as an increase in the current demand or an increase in the power demand of the drive system.

[0096] An "increase in current demand" of the drive system is a direct indication of an increase in the load. This can be measured by monitoring the current flowing through the drive system, in particular by the motor controller. An increase in current consumption indicates a greater force that is needed to run the motor. An "increase in power demand" is an indicator of an increase in load on a drive system. The power demand can be derived from the current demand and the voltage on the drive system.

[0097] According to a further or alternative embodiment, the self-driving vehicle (SDV) is configured to interpret force feedback based on contextual information originating from the state model of the programmable logic controller (PLC) and perception data from other sensors, preferably laser sensors. This contextual interpretation allows the self-driving vehicle not only to detect physical interactions, but also to evaluate them in light of the current application, the situational context, and the task to be performed. Preferably, the received force feedback is only considered relevant when it coincides with an application-specific state of the PLC model, such as, for example, performing a loading or unloading maneuver, and when simultaneous sensor perception, such as a decrease in distance from the laser sensors, confirms this interaction. This approach allows for a unique and task-oriented interpretation of force-sensor data, which distinguishes it from conventional force feedback applications. This allows the self-driving vehicle to react proactively and purposefully to physical interactions, thereby increasing both safety and precision.

[0098] According to some embodiments, force feedback is used to detect a first contact point with a side wall of a cargo space, and a second contact point with a rear wall or an already placed pallet. These contact points are respectively designated as a first and second setpoint. Preferably, measuring these setpoints allows for optimal positioning of the load within the cargo space to be achieved. After both setpoints have been reached, the vehicle is positioned such that it and the load in their longitudinal direction are substantially parallel to the side wall of the cargo space.

[0099] According to a further or alternative embodiment, the SDV comprises a loading strategy wherein the vehicle sequentially places three pallets within a cargo space. The first pallet is placed against a side wall by tilting the vehicle and pressing the pallet diagonally into the corner until contact is made, preferably detected via force feedback. The second pallet is similarly placed against the first. The third pallet is loaded into the remaining central space without a tilting movement.

[0100] According to a further or alternative embodiment, the SDV comprises an unloading strategy wherein the vehicle drives into the cargo space, detects the first pallet using the pallet camera, and subsequently orients itself with respect to this pallet. Preferably, the blocks on the underside are counted using the LiDAR sensors when driving into the pallet. Each pallet preferably consists of three rows of blocks. Upon detection of three blocks, the depth of the pallet is considered correct and the load is lifted. This procedure ensures accurate pickup of the load, even if it is not positioned against the rear of the vehicle.

[0101] The various embodiments as described herein improve the reliability, safety, and autonomy of the self-driving vehicle (SDV) in demanding logistics environments, in particular during loading and unloading maneuvers in trucks, and contribute to an optimized integration within automated warehouse infrastructures.

[0102] According to a further or alternative embodiment, the self-driving vehicle (SDV) is designed and / or configured such that its operation is completely independent of infrastructure- or building-bound systems or interactions. Preferably, the SDV requires no external markings, guidance strips, reflectors, beacons, or other fixed infrastructure elements to function correctly. The navigation, localization, obstacle detection, and control are preferably performed exclusively based on the integrated sensors and processing modules of the vehicle itself, such as the localization camera, laser sensors, ultrasonic sensors, three-dimensional (3D) cameras, and / or the industrial computer unit. This infrastructure-free operation increases the flexibility and scalability of the system and enables rapid implementation in various environments without modification of the existing infrastructure.

[0103] According to a further or alternative embodiment, the self-driving vehicle (SDV) includes a communication module, which communication module is responsible for the communication between the SDV, in particular the industrial computer, and one or more external elements such as a warehouse management system (WMS), a warehouse control system (WCS), a multi-agent system (MAS), or combinations thereof. Preferably, said communication module is based on Wi-Fi.

[0104] A "WMS" is a software application that manages the day-to-day operations of a warehouse. It helps track inventory levels, manage storage locations, process orders, and coordinate the shipping and receiving of goods. A WMS optimizes storage space and ensures a more efficient workflow within the warehouse.

[0105] The term "WCS" refers to a software application that manages and controls the physical flow of products in a warehouse. It often works in conjunction with a WMS and handles the actual movement of products through the warehouse, including the control of material handling equipment such as conveyor belts, cranes and sorting systems. The WCS ensures the real-time execution of tasks that are planned by the WMS.

[0106] A "MAS" is a system in which multiple software agents collaborate to perform tasks and solve problems. In the context of self-driving vehicles (SDVs) within a warehouse, the MAS coordinates the actions of a fleet of SDVs. The MAS receives orders from the WMS / WCS and distributes them to the SDVs. Each self-driving vehicle (SDV) operates autonomously, but the MAS ensures a coordinated and efficient execution of tasks, allowing the entire system to function optimally. The MAS ensures that the SDVs work together seamlessly and any conflicts or collisions are avoided.

[0107] According to a further or alternate embodiment, the multi-agent system (MAS) is configured to deterministically avoid conflicts and collisions between multiple autonomous vehicles (SDVs). Preferably, this is achieved through double verification in both task planning and path planning. In a first phase, potential conflicts are eliminated during the coordination of tasks and the allocation of transport commands to specific SDVs. In a second phase, path overlaps and intersections of paths between SDVs are also evaluated and excluded when generating individual path plans. Through this layered approach, it is preferably ensured that under normal circumstances, collisions and deadlocks are deterministically excluded, without the need for ad-hoc rescheduling or reactive interventions during execution. This increases the reliability and predictability of the system, particularly in complex logistics environments with multiple vehicles operating simultaneously.

[0108] According to a further or alternative embodiment, the load carrier comprises one or more forks. These forks are particularly suitable for carrying and transporting pallets. Preferably, the load carrier includes two forks. The forks are firmly attached to the vehicle base and are robust enough to be suitable for safely lifting and moving heavy loads. The forks are positioned in such a way that they can easily slide under a pallet, making the picking up and setting down of pallets efficient.

[0109] The robustness and position of the forks ensure the safe and stable movement of heavy loads, which is crucial in a warehouse environment where quickly and safely moving goods is essential. According to a further or alternative embodiment, the self-driving vehicle (SDV) includes at least three wheels. These wheels provide the self-driving vehicle with optimal stability and maneuverability. The wheels are preferably designed such that they ensure smooth and quiet operation, even on uneven surfaces. Preferably, the self-driving vehicle comprises four wheels, two of which are driven and two steered. This ensures precise control and efficient propulsion of the self-driving vehicle.

[0110] The benefits of this configuration include excellent stability and control of the vehicle, which contributes to safe and efficient operation in various environments.

[0111] According to a further or alternative embodiment, the self-driving vehicle (SDV) includes a drive system, which preferably consists of one or more electric motors. Electric motors provide precise and energy-efficient propulsion, allowing the vehicle to navigate and maneuver with high accuracy. According to some embodiments, the drive system is electronically coupled with the industrial computer unit and / or the programmable logic controller (PLC). This coupling allows the speed and direction of the self-driving vehicle to be accurately controlled. The precision and control offered by this drive system contribute to improved operational efficiency and safety.

[0112] More preferably, the drive system is electronically coupled with the programmable logic controller (PLC) via a controller area network (CAN) link.

[0113] In the context of the present invention, "controller area network (CAN)" refers to a robust vehicle bus system designed to allow microcontrollers and devices to communicate with each other without the need for a host computer. It is an international standard (ISO 11898) that was developed to enable communication between different electronic systems in vehicles.

[0114] According to a further or alternative embodiment, the self-driving vehicle (SDV) includes a lifting mechanism that allows the load carrier to lift loads from the ground. This lifting mechanism is designed to lift the load carrier, in particular the forks, vertically. A few centimeters are usually sufficient to move the pallets safely without damaging them. Preferably, the lifting mechanism is configured for lifting the load carrier over a distance comprised between 1 and 15 cm, preferably between 2 and 15 cm, more preferably between 5 and 12 cm, most preferably between 9 and 11 cm. According to some embodiments, the lifting mechanism is electronically coupled with the industrial computer unit and / or the programmable logic controller (PLC). This coupling allows for precise lifting. Preferably, the lifting mechanism is electronically linked to the programmable logic controller (PLC) by means of a controller area network (CAN) connection.

[0115] This lifting mechanism enables the self-driving vehicle to safely and efficiently lift and move pallets. It minimizes the risk of damage to the goods and ensures stable movement, which is essential for a reliable logistics operation.

[0116] According to a further or alternative embodiment, the lifting mechanism of the selfdriving vehicle (SDV) is provided with a detection system for the lifting level (lift safety level) and for the load state (load state detection). Preferably, the current position of the lifting mechanism is continuously monitored, where it is checked whether the lifting range is within predetermined safe limits, in order to avoid damage to the surroundings or to the vehicle itself. At the same time, the system detects whether there is a load on the forks or the platform, as well as whether this load is correctly positioned. This information is preferably also integrated into the PLC state model and / or transferred to the multi-agent system (MAS) for contextual decisionmaking. In this way, a safe, stable, and controlled movement of goods is guaranteed, which contributes to a reliable and damage-free logistics operation.

[0117] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises an automatic loading and unloading function, wherein the load carrier is equipped with sensors configured for detecting the presence and position of pallets. These sensors ensure that the vehicle can accurately position the forks under the pallets and safely pick up and set down the pallets. According to some embodiments, these sensors are electronically coupled with the industrial computer unit and / or the programmable logic controller (PLC) in order to ensure a seamless and efficient loading and unloading operation.

[0118] According to some embodiments, the self-driving vehicle (SDV) includes one or more indicator lights and / or buttons, which inform a user about the status of the vehicle and which may allow the user to manually control the vehicle. Preferably, said indicator lamps and / or buttons are electronically coupled with the programmable logic controller (PLC). According to a further or alternative embodiment, the self-driving vehicle (SDV) is designed and / or configured in accordance with the requirements set by the applicable CE certification guidelines regarding mobile autonomous means of transportation. In particular, it is provided that the integration of safety PLC, redundant sensors, force sensors, and force feedback, and / or the dynamic adjustment of safety fields ensure the required functional safety and risk reduction as required under Directive 2006 / 42 / EC (Machinery Directive), EN ISO 13849-1, and EN ISO 3691-4:2023.

[0119] In a second aspect, the present invention relates to a system for controlling a selfdriving vehicle (SDV) comprising: at least one self-driving vehicle (SDV), and a software architecture comprising a user interface (UI), a multi-agent system (MAS), a warehouse management system (WMS), and a warehouse control system (WCS), wherein the self-driving vehicle includes a localization camera and a plurality of safety sensors, wherein the safety sensors comprise two or more laser sensors, two or more ultrasonic sensors, and two or more three-dimensional (3D) cameras.

[0120] Preferably, the self-driving vehicle (SDV) is a self-driving vehicle according to the first aspect of the invention.

[0121] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises at least one laser scanner, preferably a LiDAR sensor, placed in a fork tip of the load carrier. Preferably, a laser scanner, preferably a LiDAR sensor, is placed in each fork tip of the load carrier, such that the vehicle obtains detailed monitoring of the space behind the forks during backward movement. Preferably, no ultrasonic sensors are present in the fork tips in this embodiment.

[0122] Thus, an independent aspect of the invention also relates to a system for controlling a self-driving vehicle (SDV) comprising: at least one self-driving vehicle (SDV), and a software architecture comprising a user interface (UI), a multi-agent system (MAS), a warehouse management system (WMS), and a warehouse control system (WCS), wherein the self-driving vehicle includes a localization camera and a plurality of safety sensors, wherein said safety sensors comprise three or more laser sensors and two or more three-dimensional (3D) cameras. According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises an industrial computer unit and a programmable logic controller (PLC), which industrial computer unit and / or programmable logic controller (PLC) include instructions for controlling processes comprising mapping, localization, carrier detection, carrier type identification, obstacle detection, PLC communication, vehicle control, or combinations thereof. The system is designed to support parallel processes, thus ensuring efficient and reliable operation of the SDV.

[0123] In the context of the present invention, these processes are to be understood as follows.

[0124] "Mapping" is the process of mapping the environment in which the self-driving vehicle (SDV) operates. This is achieved by collecting visual and sensory data that are used to create a detailed map, allowing the SDV to plan and execute its navigation paths.

[0125] "Localization" is the process of determining the position of the self-driving vehicle (SDV) within the map of the environment. This is done by processing data from the localization camera and other sensors to determine the exact location and orientation of the vehicle in real-time.

[0126] "Carrier detection" is the process of identifying and locating objects to be transported, such as pallets. This includes the use of sensors and cameras to detect objects and determine their position relative to the self-driving vehicle (SDV), so that the vehicle can safely and efficiently pick up and move these objects.

[0127] "Obstacle detection" is the process of detecting objects and people in the immediate vicinity of the self-driving vehicle (SDV). This is achieved through the use of a plurality of sensors, including laser sensors, ultrasonic sensors, and three- dimensional (3D) cameras, which work together to quickly and accurately identify and localize obstacles.

[0128] "PLC communication" is the process of communication between the programmable logic controller (PLC) and the rest of the autonomous vehicle (SDV) hardware and software. This includes sending commands from the PLC to the actuators and receiving feedback from sensors to coordinate and control the operation of the vehicle. "Vehicle control" includes the processes responsible for steering, accelerating, decelerating and stopping the hardware and software (SDV). This includes the execution of navigation paths, obstacle avoidance, picking up and setting down pallets, and other critical functions required for the autonomous operation of the vehicle.

[0129] These parallel processes ensure fast and efficient processing of sensor data and execution of vehicle actions. The structured software architecture makes it easier to expand and adapt functionalities to new requirements or technologies. By grouping the processes and allowing them to work in parallel, the overall reliability and performance of the SDV is improved, resulting in more robust operations.

[0130] According to a further or alternative embodiment, the mapping process of the selfdriving vehicle (SDV) comprises the following steps:

[0131] (i) establishing an origin point for the start of the mapping process in a space to be mapped;

[0132] (ii) navigating the self-driving vehicle (SDV) along a mapping loop in the space to be mapped, which mapping loop starts at the origin point from step (i);

[0133] (iii) navigating the self-driving vehicle (SDV) along a subsequent mapping loop in the space to be mapped, which subsequent mapping loop at least partially overlaps with the preceding mapping loop;

[0134] (iv) repeating step (iii) until an intended area in the space to be mapped has been mapped; and

[0135] (v) combining the mapping loops.

[0136] The term "space to be mapped" refers to the specific area or environment in which the self-driving vehicle (SDV) will operate and for which a detailed map must be created. This space comprises all locations and routes where the SDV will drive and requires accurate and complete mapping to enable safe and efficient navigation.

[0137] The term "mapping loop" refers to a specific driving route that the self-driving vehicle (SDV) follows during the mapping procedure. Each mapping loop is a closed route that partially overlaps with the previous loop, whereby the collected data can be combined to create a consistent and detailed map of the space to be mapped. Preferably, the mapping process further comprises the step of: (vi) performing a "pose rectification." The "pose rectification" allows any differences in the z-axis (height) to be eliminated. This ensures a uniform and consistent map of the entire environment.

[0138] According to a further or alternative embodiment, the localization process of the selfdriving vehicle (SDV) is performed using a combination of data obtained from the localization camera and the plurality of safety sensors. Preferably, the localization process allows for an accurate determination of the position and orientation of the self-driving vehicle in real-time.

[0139] According to a further or alternative embodiment, the localization process of the selfdriving vehicle (SDV) uses a Kalman filter. Preferably, the Kalman filter combines the data obtained from the localization camera and the plurality of safety sensors. The use of the Kalman filter helps in reducing noise and correcting any errors in the sensor data, which results in an accurate estimate of the location and orientation of the selfdriving vehicle.

[0140] In the context of the present invention, the term "Kalman filter" refers to an algorithm that is used to process and combine a series of measurement data over time, whereby noise and other uncertainties are reduced to make an accurate estimate of variables such as position and speed of the self-driving vehicle.

[0141] According to a further or alternative embodiment, the localization process of the selfdriving vehicle (SDV) uses Simultaneous Localization and Mapping (SLAM). Preferably, SLAM creates a dynamic map of the environment while the self-driving vehicle is in motion. SLAM technology allows the self-driving vehicle to determine its own position and simultaneously build a map of its environment.

[0142] This has the advantage that the self-driving vehicle is able to accurately navigate and operate in complex and dynamic environments, such as warehouses and production facilities, where the presence of people and unexpected obstacles poses a risk. Through accurate localization, the vehicle can operate more efficiently and safely, which leads to improved operational efficiency and increased safety for people in the vicinity.

[0143] According to a further or alternative embodiment, the "carrier detection" process comprises the detection of a carrier, or pallet, when a pallet is present behind the self-driving vehicle (SDV). This detection is performed using a pallet camera, preferably a LiDAR camera.

[0144] According to a further or alternative embodiment, the detection of a pallet comprises:

[0145] (i) detecting a pallet point cloud, preferably by observing a pallet with the pallet camera;

[0146] (ii) analyzing and / or evaluating a possible correspondence between said plurality of points and a predefined pallet model; and

[0147] (iii) if a correspondence is found in step (ii), determining a location of the pallet relative to the self-driving vehicle.

[0148] A "pallet point cloud" indicates a plurality of points which are detected when observing a pallet, preferably by means of the pallet camera.

[0149] Preferably, step (ii) is performed by means of a Cloud Match Server. A "Cloud Match Server" in the context of the present invention includes algorithms to analyze and compare point cloud data, i.e., the detected plurality of points, from the pallet camera with a predefined pallet model, for example an EU-pallet. If correspondences are found, this gives rise to a so-called match.

[0150] The advantage of this detection and matching procedure is that the SDV can accurately locate, load and / or unload pallets, even in complex and busy environments. The accurate detection and localization of pallets improve the efficiency of loading and unloading operations, which is essential for a smooth and effective operation of warehouse and logistics processes.

[0151] Thanks to this accurate detection and matching procedure, the SDV can safely and efficiently pick up and move pallets, which leads to improved operational efficiency and increased safety for the goods and the environment in which the SDV operates.

[0152] The detection and matching procedure as described herein is primarily relevant for detecting loads, preferably pallets, in the vicinity of the self-driving vehicle. However, as soon as the self-driving vehicle proceeds to pick up these loads, preferably pallets, the type of pallet is also relevant. Thus, the system according to the present invention may comprise a carrier type identification process. Carrier type identification" concerns the process wherein a type of pallet is identified just before lifting the pallet and / or during lifting of the pallet.

[0153] According to a further or alternative embodiment, the carrier type identification process comprises identifying a type of pallet by means of said laser scanners. Preferably, the carrier type identification process comprises identifying a type of pallet by means of monitoring one or more load detection fields by said laser scanners.

[0154] Preferably, the one or more load detection fields are substantially directed towards the rear of the vehicle, more preferably towards a load on the load carrier, more preferably towards a pallet on the load carrier.

[0155] According to some embodiments, the type of pallet comprises a Euro-pallet or an industrial pallet. Euro-pallets have standard dimensions of 800 x 1000 mm, and industrial pallets have standard dimensions of 1000 x 1200 mm.

[0156] According to some embodiments, the one or more load detection fields are designed for detecting and identifying Euro-pallets, preferably within the width of the vehicle base, and / or for detecting and identifying industrial pallets, preferably with a margin of 10 cm.

[0157] According to some embodiments, monitoring one or more load detection fields comprises monitoring a first load detection field that is designed for detecting and identifying Euro-pallets, preferably within the width of the vehicle base, and monitoring a second load detection field that is designed for detecting and identifying industrial pallets, preferably with a margin of 10 cm. Each of said laser scanners in the system is thus individually capable of detecting whether the present pallet is a Euro-pallet or an industrial pallet. Thus, the laser scanners are able to verify with respect to each other whether a correct type of pallet was identified.

[0158] Preferably, said laser scanners switch repeatedly between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill. More preferably, said laser scanners repeatedly switch between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill at a frequency comprised between 1 and 2 Hz, more preferably between 1.2 and 2 Hz, between 1.4 and 2 Hz, between 1.6 and 2 Hz, or between 1.8 and 2 Hz. Most preferably, said laser scanners repeatedly switch between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill at a frequency of 2 Hz. This rapid switching allows load detection to take place efficiently just before lifting a pallet and / or during the lifting of a pallet.

[0159] According to some embodiments, said laser scanners switch at least 5 times between the first load detection field and the second load detection field when the self-driving vehicle is at a standstill, preferably at least 10 times, more preferably at least 15 times, or most preferably at least 20 times.

[0160] According to a further or alternative embodiment, the safety PLC is configured to adapt the safety fields based on a type of pallet identified by said laser scanners. In particular, when an industrial pallet is identified, the safety fields are enlarged to compensate for the larger dimensions of the pallet. This prevents hazardous situations where incorrect safety fields are used.

[0161] According to some embodiments, the identification of a pallet type by said laser scanners meets performance level PL-d when the self-driving vehicle is at a standstill, and performance level PL-b when the self-driving vehicle is in motion.

[0162] According to a further or alternative embodiment, the obstacle detection process uses data obtained from the safety sensors and the safety PLC. The obstacle detection process is designed to avoid obstacles in a more intelligent manner, and thus possibly avoid stopping the SDV.

[0163] According to a further or alternative embodiment, the self-driving vehicle comprises an SDV manager. In particular, when an obstacle is detected in front of the SDV, the SDV manager checks whether this obstacle is moving or not, and / or whether another path can be calculated to avoid this obstacle. This intelligence is not implemented in the obstacle detection process itself, but in the SDV manager.

[0164] In the context of the present invention, the term "SDV manager" refers to a central coordination unit within the self-driving vehicle (SDV) that is responsible for performing actions, publishing paths, and monitoring the status of the performed actions. According to a further or alternative embodiment, said data obtained from the safety sensors comprises the data from the laser sensors, the ultrasonic sensors, and the three-dimensional (3D) cameras. Preferably, the data from these sensors are grouped by a Pointcloud Smasher, whereby an obstacle point cloud is obtained.

[0165] An "obstacle point cloud" indicates a plurality of points which are detected when observing an obstacle, preferably by means of the safety sensors.

[0166] The obstacle detection process preferably analyzes the detected obstacle in one or more facets chosen from the obstacle's state of motion, the obstacle's size, the obstacle's speed, the obstacle's direction, possible collision with the SDV, or combinations thereof.

[0167] The information obtained from said facets is used by the SDV manager to make decisions about avoiding obstacles.

[0168] The PLC communication process is, according to some embodiments, responsible for the communication between the programmable logic controller (PLC) and the software instructions of the self-driving vehicle (SDV). According to some embodiments, the PLC communication process reads one or more parameters chosen from the group of traction current, lifting status, odometry, selected monitoring fields, wheel speed, carrier weight, battery status, PLC status, or combinations thereof. According to some embodiments, the PLC communication process sends one or more parameters to the PLC chosen from the group of desired lifting status, desired speed, desired PLC status, or combinations thereof.

[0169] In the context of the present invention, the term "traction current" refers to the electrical current used by the motors of the self-driving vehicle (SDV) for propulsion.

[0170] The term "lifting status" indicates the current position or condition of the lift mechanism.

[0171] The term "odometry" refers to the collected data relating to the distance traveled and position of the SDV over time.

[0172] The term "selected monitoring fields" indicates the specific areas around the SDV that are monitored by the safety sensors to detect obstacles and hazards. The term "wheel speed" refers to the rotational speed of the wheels of the SDV.

[0173] "Carrier weight" means the weight of the load being transported by the load carrier.

[0174] "Battery status" provides information about the charge level and performance of the SDV battery.

[0175] The "PLC status" pertains to the operational state and functional condition of the PLC.

[0176] The PLC communication process has the advantage that an efficient and reliable data exchange can take place between the PLC and the SDV software. By reading realtime data from the PLC and sending the required commands, the SDV can be accurately and responsively controlled, which leads to improved operational efficiency and safety.

[0177] According to some embodiments, the vehicle control process comprises one or more sub-processes chosen from the group of interacting with the SDV manager, interacting with a flexible behavior module, interacting with a path planning module, interacting with a path execution module, or combinations thereof.

[0178] The SDV manager is responsible for reading the desired actions to be performed by the SDV originating from the multi-agent system (MAS). Subsequently, the SDV manager publishes paths to be executed, desired PLC status, and / or desired behaviors of the self-driving vehicle (SDV). The SDV manager also reads the status of these behaviors and checks if everything is proceeding as expected.

[0179] The "flexible behavior module" is an intelligent module in which all complex behaviors are programmed. Examples of this are how a carrier should be picked up or how an obstacle should be avoided.

[0180] The "path planning module" is responsible for calculating paths between two positions. This module is used, for example, when the SDV is in front of a carrier to calculate a path in order to pick up that carrier. The "path execution module" is responsible for monitoring the execution of the desired paths. This module generates the speed commands that are sent and monitors the correct execution of the paths.

[0181] The advantage of these integrated processes and / or modules is that they ensure an advanced and reliable operation of the SDV. The combination of the SDV manager, flexible behavior module, path planning module, and path execution module ensures that the vehicle is able to perform complex tasks, avoid obstacles, and navigate efficiently, which results in increased operational efficiency and safety.

[0182] According to a further or alternative embodiment, the user interface (UI) comprises elements with which specific actions can be requested and monitored, including mapping, calibration, and configuration of the SDV. The UI communicates with the software components of the SDV and enables users to interact with the vehicle via a web browser.

[0183] The UI offers user-friendliness, allowing users to easily interact with the SDV, even remotely. Easy access from virtually any device with an internet connection increases flexibility and accessibility. The ability to perform mapping, calibration, and configuration via the UI, makes managing the SDV simpler and faster, which saves time and resources. Preferably, the user interface (UI) is a web-based user interface.

[0184] According to a further or alternative embodiment, the multi-agent system (MAS) is configured for coordinating one or more SDVs in a fleet. Preferably, the MAS receives orders from a WMS and / or WCS and distributes these among the SDVs via a standard communication protocol, such as the VDA5050 protocol. More preferably, each SDV operates autonomously, and the MAS ensures coordinated and efficient execution of tasks.

[0185] This has the advantage that the MAS optimizes the task allocation among the SDVs, which leads to increased operational efficiency and a reduction in the time required to complete logistics tasks. By using a standard communication protocol, SDVs from different manufacturers can seamlessly cooperate within the same system. The central coordination by the MAS reduces the risk of collisions and ensures a smooth and efficient workflow in warehouse environments. According to a further or alternative embodiment, the system comprises a communication infrastructure that ensures the SDV receives orders from the WMS, WCS, and / or MAS, preferably via Wi-Fi or Ethernet, most preferably via Wi-Fi. The communication infrastructure ensures that the industrial computer unit of the SDV receives, processes, and transmits these orders to the PLC, and possibly the safety PLC, for execution.

[0186] This has the advantage that the communication infrastructure ensures a reliable and fast transfer of orders and commands between the various systems and the SDV. This results in seamless integration and coordination between the software components, which leads to efficient and effective execution of logistics tasks. The use of Wi-Fi or Ethernet offers flexibility in the implementation of the system in various operational environments.

[0187] Specific embodiments relating to the first and / or second aspect of the invention are as follows.

[0188] According to a further or alternative embodiment, the self-driving vehicle (SDV) is constructed with a view to robustness and industrial deployability in warehouse environments without the need for adapted infrastructure. To this end, the vehicle comprises standard components that are known from conventional manual forklift trucks, including motors, drive wheels, support wheels, and fork wheels. The forks are preferably interchangeable, and can, depending on the application, be replaced by short or long fork versions.

[0189] Preferably, the vehicle is provided with a specific lifting system, configured to keep the SLAM camera continuously horizontal during lifting movements.

[0190] According to a further or alternative embodiment, the vehicle comprises a compact chassis design, without a fork-over configuration, whereby the vehicle is compatible with industrial pallets. The lifting mechanism is preferably a pull-rod configuration and allows a lifting height of preferably 120 mm, wherein the system only has two fixed positions: fully up or fully down. The maximum lifting capacity is preferably 1500 kg.

[0191] According to a further or other embodiment, the SDV is provided with a battery type based on lithium titanate oxide (LTO), which battery comprises a nominal voltage of 24V and a capacity of preferably 65 Ah. Owing to this technology, the operating time of the vehicle is preferably 2.5 to 3 hours, with a charging time of 15 minutes from 10% to 90% via an automatic docking system coupled to a 3-phase charger (400 V AC to 24 V DC, 300 A charging current).

[0192] According to a further or other embodiment, the SDV is provided with an industrial computer unit based on an Intel Core i7-10700TE processor with a minimum of 32 GB RAM and 500 GB storage, running on an Ubuntu 18.04.06 LTS operating system. The industrial PC is configured for Wi-Fi connectivity on 2.4 GHz and 5 GHz networks.

[0193] According to a further or other embodiment, the SDV uses 3D visual SLAM technology for localization, wherein no physical infrastructure is required. The system supports autonomous or semi-autonomous mapping functionality with a positioning accuracy of ±3 cm and an orientation accuracy of ±1°. The detection of pallets has a similar accuracy. Preferably, this is a positioning accuracy of ±3 cm at the fleet level. More preferably, this is a positioning accuracy of ±2 cm at the vehicle level.

[0194] According to a further or other embodiment, the maximum forward driving speed of the SDV is comprised between 2.0 m / s and 3.0 m / s. According to some embodiments, the maximum reverse driving speed is comprised between 0.3 and 1.5 m / s. Preferably, the maximum forward driving speed is comprised between 2.5 m / s and 3.0 m / s. Even more preferably, the maximum reverse driving speed is comprised between 0.6 and 1.5 m / s, more preferably comprised between 1.0 and 1.5 m / s. The minimum required aisle width varies from 1400 mm or less at 0.4 m / s to 2000 mm or less at 2.0 m / s. Preferably, the minimum required aisle width is 1300 mm or less at 0.4 m / s, more preferably the minimum required aisle width is 1200 mm or less at 0.4 m / s. According to some embodiments, the minimum required aisle width is 1900 mm or less at 2.0 m / s. Preferably, the minimum required aisle width is 1800 mm or less at 2.0 m / s, more preferably the minimum required aisle width is 1700 mm or less at 2.0 m / s. In specific circumstances wherein the safety scanners are muted, this can be reduced to 900 mm, preferably less than 900 mm.

[0195] According to a further or other embodiment, the SDV is controlled via ROS (Robot Operating System) as middleware in combination with a Multi-Agent System (MAS) that organizes communication via the VDA5050 protocol. The MAS coordinates, inter alia, automatic task allocation, fleet management, traffic rules, loading updates, and intelligent battery management. According to a further or other embodiment, the SDV comprises a web-based user interface for individual vehicles, separate from the MAS interface. This user interface provides functionality for layout design, setting waypoints and driving paths, configuring loading and unloading locations, and defining parking and charging zones.

[0196] According to a further or other embodiment, the vehicle is capable of detecting obstacles over the entire width and height of the chassis, and is equipped with path replanning to dynamically avoid these obstacles. Pallet detection is possible with lateral or axial deviations up to ±20 cm and with a relative angular rotation between the vehicle and the pallet of a maximum of 30°.

[0197] A third aspect of the present invention concerns a method for controlling a self-driving vehicle (SDV) comprising the steps of:

[0198] (a) obtaining a command for a self-driving vehicle (SDV), said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base;

[0199] (b) localizing said self-driving vehicle;

[0200] (c) optionally, detecting one or more obstacles around the self-driving vehicle;

[0201] (d) based on the command obtained in step (a), a location of the self-driving vehicle obtained in step (b), and optionally, a location of said obstacles around the self-driving vehicle obtained from step (c), generating a navigation path for the self-driving vehicle; and

[0202] (e) executing the navigation path generated in step (d), wherein steps (b) and (c) are performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising two or more laser sensors, two or more ultrasonic sensors, and two or more three-dimensional (3D) cameras.

[0203] According to some embodiments, the command from step (a) is a navigation command.

[0204] A "navigation command" is to be understood as a command wherein the self-driving vehicle is moved from the location obtained from step (b) to a target location.

[0205] According to some embodiments, the command from step (a) is a loading and / or unloading command. A "loading and / or unloading command" is to be understood as a command wherein the self-driving vehicle loads, or picks up, a load, and / or unloads, or sets down, a load.

[0206] According to some embodiments, the command is a loading command, and the method for this purpose comprises the steps of:

[0207] (f) detecting a load in the vicinity of the self-driving vehicle,

[0208] (g) orienting the self-driving vehicle in an orientation that allows loading, or picking up, the load,

[0209] (h) loading, or picking up, the load by maneuvering the self-driving vehicle, wherein the load carrier is moved under the load, and subsequently

[0210] (i) lifting the load carrier.

[0211] According to some embodiments, the command is an unloading command, and the method for this purpose comprises the steps of:

[0212] (j) unloading, or setting down, the load by lowering the load carrier, and subsequently

[0213] (k) maneuvering the self-driving vehicle, wherein the load carrier is moved out from under the load.

[0214] According to a preferred embodiment, the self-driving vehicle (SDV) is a self-driving vehicle according to the first aspect of the invention.

[0215] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises at least one laser scanner, preferably a LiDAR sensor, placed in a fork tip of the load carrier. Preferably, a laser scanner, preferably a LiDAR sensor, is placed in each fork tip of the load carrier, such that the vehicle obtains detailed monitoring of the space behind the forks during backward movement. Preferably, no ultrasonic sensors are present in the fork tips in this embodiment.

[0216] Thus, an independent aspect of the invention also concerns a method for controlling a self-driving vehicle (SDV) comprising the steps of:

[0217] (a) obtaining a command for a self-driving vehicle (SDV), said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base;

[0218] (b) localizing said self-driving vehicle;

[0219] (c) optionally, detecting one or more obstacles around the self-driving vehicle; (d) based on the command obtained in step (a), a location of the self-driving vehicle obtained in step (b), and optionally, a location of said obstacles around the self-driving vehicle obtained from step (c), generating a navigation path for the self-driving vehicle; and

[0220] (e) executing the navigation path generated in step (d), wherein steps (b) and (c) are performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising three or more laser sensors and two or more three-dimensional (3D) cameras. According to a further or other preferred embodiment, the method uses a system according to the second aspect of the invention.

[0221] According to a further or other embodiment, step (c) further comprises the step of: combining data obtained from the two or more laser sensors, the two or more ultrasonic sensors, and the two or more three-dimensional (3D) cameras into an obstacle point cloud.

[0222] According to a further or other embodiment, step (c) further comprises the step of: combining data obtained from the safety sensors to define a monitoring field, which monitoring field is at least 10% larger than a safety field around the self-driving vehicle.

[0223] The method as described herein has the advantage that the various sensors of the self-driving vehicle not only monitor the direct safety field, but also scan a larger area within the field of view, namely the monitoring field. This enables the system to detect potential obstacles and hazards at an early stage and thus proactively adjust the control of the vehicle, which can reduce the need for sudden emergency stops.

[0224] Preferably, the monitoring field is at least 15% larger than a safety field around the self-driving vehicle, more preferably, the monitoring field is at least 20% larger than the safety field around the self-driving vehicle, even more preferably, the monitoring field is at least 25% larger than the safety field.

[0225] Proactively adjusting the control of the self-driving vehicle based on monitoring of the monitoring field, comprises, according to some embodiments, deviating from a planned path, decelerating, accelerating, and / or bringing the self-driving vehicle to a stop. According to a further or other embodiment, steps (d) and / or (e) further comprise: (i) adjusting the navigation path based on the location of said obstacles around the self-driving vehicle (SDV), and / or (ii) the self-driving vehicle deviating from the generated navigation path based on the location of said obstacles.

[0226] Adjusting the navigation path and / or deviating from the generated navigation path allow the SDV to continue driving at a higher speed, whereby decelerations and / or emergency stops and restarts are avoided.

[0227] According to some embodiments, the method comprises one or more steps for performing one or more processes comprising mapping, localization, carrier detection, carrier type identification, obstacle detection, PLC communication, vehicle control, or combinations thereof.

[0228] A fourth aspect of the present invention concerns a method for loading a truck by means of a self-driving vehicle (SDV) comprising the steps of:

[0229] (a) navigating a self-driving vehicle (SDV) up to a truck, preferably to the center of a tailgate of said truck, said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base, wherein the load carrier is loaded with a load, and which truck comprises a cargo space;

[0230] (b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of the cargo space;

[0231] (c) navigating the self-driving vehicle into the cargo space of the truck, until the self-driving vehicle has reached a suitable location for unloading the load;

[0232] (d) unloading the load from the load carrier of the self-driving vehicle into the cargo space of the truck; and

[0233] (e) navigating the self-driving vehicle out of the cargo space of the truck, preferably by navigating to the center of the cargo space of the truck, and then navigating away from the cargo space; wherein the navigating and / or maneuvering of the self-driving vehicle is performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising two or more laser sensors, two or more ultrasonic sensors, and two or more three-dimensional (3D) cameras. The method has the advantage that the combination of sensors allows the self-driving vehicle (SDV) to efficiently and controllably navigate and / or maneuver both into and out of the cargo space of the truck.

[0234] According to a preferred embodiment, the self-driving vehicle (SDV) is a self-driving vehicle according to the first aspect of the invention.

[0235] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises at least one laser scanner, preferably a LiDAR sensor, placed in a fork tip of the load carrier. Preferably, a laser scanner, preferably a LiDAR sensor, is placed in each fork tip of the load carrier, such that the vehicle obtains detailed monitoring of the space behind the forks during backward movement. Preferably, no ultrasonic sensors are present in the fork tips in this embodiment.

[0236] Accordingly, an independent aspect of the invention also relates to a method for loading a truck by means of a self-driving vehicle (SDV) comprising the steps of:

[0237] (a) navigating a self-driving vehicle (SDV) up to a truck, preferably to the center of a tailgate of said truck, said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base, wherein the load carrier is loaded with a load, and which truck comprises a cargo space;

[0238] (b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of the cargo space;

[0239] (c) navigating the self-driving vehicle into the cargo space of the truck, until the self-driving vehicle has reached a suitable location for unloading the load;

[0240] (d) unloading the load from the load carrier of the self-driving vehicle into the cargo space of the truck; and

[0241] (e) navigating the self-driving vehicle out of the cargo space of the truck, preferably by navigating to the center of the cargo space of the truck, and then navigating away from the cargo space; wherein the navigating and / or maneuvering of the self-driving vehicle is performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising three or more laser sensors and two or more three-dimensional (3D) cameras. According to a further or other preferred embodiment, the method uses a system according to the second aspect of the invention.

[0242] According to a further or other embodiment of the fourth aspect of the present invention, the method comprises additional preparatory steps prior to loading the truck, wherein said preparatory steps comprise:

[0243] (i) navigating the self-driving vehicle (SDV) to a shipping lane in which a load is presented;

[0244] (ii) positioning the SDV relative to the presented load in the shipping lane;

[0245] (iii) picking up the load onto the load carrier of the SDV; wherein the method is then continued by navigating the SDV to the truck in accordance with step (a) of the fourth aspect.

[0246] According to a further or other embodiment, step (a) is preceded by the step of analyzing the cargo space of the truck. This step is preferably performed by means of the localization camera and / or the plurality of safety sensors. This step allows to check whether loads, preferably pallets, are already present in the cargo space, and if so, where they are positioned. Based on this, the self-driving vehicle can better decide where pallets should be loaded in the cargo space of the truck.

[0247] According to a further or other embodiment, the method comprises, between steps (a) and (b) and / or between steps (b) and (c), the step of analyzing one or more walls of the cargo space of the truck. This step is preferably performed by means of the localization camera and / or the plurality of safety sensors. This step allows the selfdriving vehicle to orient and / or position itself in the cargo space relative to the walls of the cargo space.

[0248] According to a further or other preferred embodiment, the method in step (a) uses the steps as described in the method according to the third aspect of the invention.

[0249] According to a further or other embodiment, step (c) is performed by means of the two or more laser sensors.

[0250] Navigation within a cargo space of a truck is particularly challenging as the cargo space is a dark environment, in which light-sensitive sensors such as cameras operate suboptimally. Simultaneous Localization and Mapping (SLAM) can therefore not be optimally used for navigating the self-driving vehicle in the cargo space. Moreover, even if the use of Simultaneous Localization and Mapping (SLAM) were possible, the accuracy of this system is insufficient to allow for the efficient loading of a truck's cargo space. However, the use of two or more laser sensors does allow loading of the cargo space.

[0251] According to some embodiments, the cargo space of the truck offers space in width for three loads positioned next to each other, preferably pallets. More preferably, the method is performed such that first the two loads, preferably pallets, are loaded against the side walls, and then the load between the two preceding loads, preferably pallets, is loaded.

[0252] According to some embodiments, step (c) comprises the following sub-steps, particularly when a load is to be placed against a side wall:

[0253] (i) navigating the self-driving vehicle in the cargo space towards a side wall of the cargo space;

[0254] (ii) tilting the self-driving vehicle, wherein a corner of the load, preferably the pallet, is moved up against the side wall of the cargo space; and

[0255] (iii) maneuvering the self-driving vehicle until the load is positioned against the rear wall of the cargo space, or a load behind it, and / or until the selfdriving vehicle and the load are oriented substantially parallel in their longitudinal direction to the side wall of the cargo space; wherein during step (ii) the two or more laser sensors monitor the distance between the self-driving vehicle and / or the load, and the side wall of the cargo space.

[0256] The use of the laser sensors is extremely suitable for monitoring the distance in the cargo space of the truck between the self-driving vehicle and / or the load, and the side wall of the cargo space, as they function optimally in dark conditions and moreover have high accuracy.

[0257] According to a further or other embodiment, step (ii) comprises detecting one or more obstacles around the self-driving vehicle, preferably just before the corner of the load, preferably the pallet, is moved up against the side wall of the cargo space. According to a further or other embodiment, step (ii) comprises monitoring the distance between the self-driving vehicle and / or the load, and the opposite side wall of the cargo space. Performing this detection before the side wall of the cargo space is touched ensures efficient checking for the presence of persons or obstacles on the left side of the selfdriving vehicle and / or the load, and moreover verifies that the safety field on the right side covers the full width of the truck. In this way, correct positioning of the load, in particular the pallet, is obtained.

[0258] According to a further or other embodiment, step (iii) comprises scraping and / or dragging the edge of the load against the side wall of the cargo space during the maneuvering of the self-driving vehicle until the self-driving vehicle and the load are oriented substantially parallel in their longitudinal direction to the side wall of the cargo space.

[0259] Preferably, steps (ii) and / or (iii) are performed by means of force feedback from a drive system of the self-driving vehicle.

[0260] In the context of the present invention, the term "force feedback" is defined as a mechanical or electronic system that feeds back forces or resistive forces experienced by a vehicle, such as a self-driving vehicle (SDV), to the control systems of the vehicle. This enables the vehicle to react in real-time to physical interactions with its environment by making adjustments to the movement or navigation parameters.

[0261] Specifically for the self-driving vehicle (SDV) in this invention, force feedback is used to monitor and correct the interaction between the load carrier of the vehicle and the load or side walls of a cargo space. When the load carrier, for example, presses against an obstacle or the side wall of the cargo space, force feedback allows this force to be detected, and a signal to be sent back to the control systems in the SDV. This information can be optimally used to decelerate, stop, and / or adjust the movement of the self-driving vehicle, in particular the load carrier, whereby the load carrier assumes a correct position without causing damage to the load or the truck.

[0262] Force feedback contributes to the precision and safety of loading and unloading by enabling the SDV to make subtle adjustments based on the actual forces experienced.

[0263] Force feedback is, according to some embodiments, measured by means of one or more force sensors of the self-driving vehicle. Force feedback can, according to some embodiments, also be measured without separate force sensors. According to some embodiments, force feedback is measured as an increase in load on the drive system, for example, as an increase in the current demand or an increase in the power demand of the drive system.

[0264] An "increase in current demand" of the drive system is a direct indication of an increase in the load. This can be measured by monitoring the current flowing through the drive system, in particular by the motor controller. An increase in current consumption indicates a greater force that is needed to run the motor.

[0265] An "increase in power demand" is an indicator of an increase in load on a drive system. The power demand can be derived from the current demand and the voltage on the drive system.

[0266] In particular, according to some embodiments, steps (ii) and / or (iii) comprise measuring a first setpoint in force feedback. The first setpoint as described herein concerns a first increase in load that the SDV experiences when the load touches the side wall of the cargo space and / or when scraping and / or dragging the edge of the load against the side wall of the cargo space. Based on this first setpoint, the SDV is thus aware of its position against the side wall.

[0267] According to a further or other embodiment, step (iii) comprises measuring a second setpoint in force feedback. The second setpoint as described herein concerns a second increase in load that the SDV experiences when the load touches the rear wall of the cargo space, or a load behind it. Based on this second setpoint, the SDV is thus aware of its position against the rear wall.

[0268] According to some embodiments, after the first and second setpoints have been reached, the method comprises the step of maneuvering the self-driving vehicle until the self-driving vehicle and the load are oriented substantially parallel in their longitudinal direction to the side wall of the cargo space.

[0269] According to some embodiments, the safety fields of the various laser scanners can be dynamically enabled, disabled, and / or limited during different steps of loading and / or unloading.

[0270] According to some embodiments, during (i) navigating the self-driving vehicle in the cargo space towards a side wall of the cargo space, the laser scanners coupled to the front and / or the lateral sides of the vehicle are enabled, and the laser scanners in the fork tips are enabled. According to some embodiments, in step (ii), tilting the self-driving vehicle, wherein a corner of the load, preferably the pallet, is moved up against the side wall of the cargo space, the laser scanners coupled to the front side and / or the lateral sides of the vehicle are activated, and the laser scanners in the fork tips are deactivated. This allows the vehicle to approach the side wall of the cargo space closely enough without triggering safety fields.

[0271] According to some embodiments, in step (iii), maneuvering the self-driving vehicle until the load is positioned against the rear wall of the cargo space, or a load behind it, and / or until the self-driving vehicle and the load are oriented substantially parallel in their longitudinal direction to the side wall of the cargo space, the laser scanners coupled to the front side and / or the lateral sides of the vehicle are activated, and the laser scanners in the fork tips are initially activated, or else deactivated in a final phase. Thus, safety is initially ensured; however, as soon as persons can no longer be present in the remaining space in front of the vehicle, deactivating the laser scanners in the fork tips allows the load to be correctly positioned without triggering safety fields. Preferably, the laser scanners are deactivated in said final phase as soon as the remaining space in front of the vehicle is smaller than 180 mm.

[0272] According to a further or alternative embodiment, step (c) comprises the following sub-steps, in particular when a load is to be placed in the middle of the cargo space, between two already present loads which are located against the side walls of the cargo space:

[0273] (i) navigating the self-driving vehicle to the middle of the cargo space, in particular to the middle between two side walls of the cargo space; and

[0274] (ii) navigating the self-driving vehicle, whereby the load carrier loaded with the load is brought between two already present loads.

[0275] According to some embodiments, step (ii) comprises measuring a setpoint in force feedback. The setpoint as described herein concerns an increase in load that the SDV experiences when the load contacts the rear wall of the cargo space, or a load behind it.

[0276] Various embodiments of the method according to the fourth aspect of the invention have the advantage that the self-driving vehicle (SDV) can be agnostic about the depth of the cargo space of the truck. The various sensors and method steps as described herein allow the self-driving vehicle to load the truck without prior knowledge of the dimensions of its cargo space. In particular, the various embodiments as described herein allow cargo spaces, preferably cargo spaces with a width comprised between 2470 and 2540 mm, to be loaded without sacrificing accuracy and / or performance. The fact that the self-driving vehicle is capable of autonomously performing the loading operation makes integration with a WMS for loading a truck superfluous.

[0277] According to a particular embodiment, the loading of a truck is performed according to a predetermined sequence:

[0278] (1) The SDV positions the first pallet against a side wall of the cargo space, wherein the vehicle is moved diagonally until the corner of the pallet makes contact with the side wall. This contact is detected via an increase in current or power demand of the drive system (force feedback first setpoint);

[0279] (2) Subsequently, a second pallet is pressed against the already placed pallet in a similar manner;

[0280] (3) The third pallet is then placed in the remaining central space.

[0281] According to some embodiments, force feedback is also used to detect a second setpoint during positioning, which corresponds to contact between the pallet and the rear wall of the cargo space or an already placed pallet. This allows the SDV to orient itself and finally position itself with high accuracy. According to a further or other embodiment, the self-driving vehicle (SDV) is configured for force-controlled loading and unloading of pallets, preferably wherein force sensors and control algorithms are employed. This allows real-time corrections to be performed during loading and unloading based on measured physical interactions between the forks and the load or environment, in order to avoid damage to the goods or infrastructure.

[0282] In a fifth aspect, the present invention concerns a method for unloading a truck by means of a self-driving vehicle (SDV) comprising the steps of:

[0283] (a) navigating a self-driving vehicle (SDV) up to a truck, preferably to the center of a tailgate of said truck, said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base, wherein the load carrier is loaded with a load, and which truck comprises a cargo space;

[0284] (b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of the cargo space; (c) navigating the self-driving vehicle into the loading space of the truck, until the self-driving vehicle has reached a suitable location for loading the load;

[0285] (d) loading the load from the cargo space of the truck onto the load carrier of the self-driving vehicle; and

[0286] (e) navigating the self-driving vehicle out of the cargo space of the truck, preferably by navigating to the center of the cargo space of the truck, and then navigating away from the cargo space; wherein the navigating and / or maneuvering of the self-driving vehicle is performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising two or more laser sensors, two or more ultrasonic sensors, and two or more three-dimensional (3D) cameras.

[0287] The method has the advantage that the combination of sensors allows the self-driving vehicle (SDV) to efficiently and controllably navigate and / or maneuver both into and out of the cargo space of the truck.

[0288] According to a preferred embodiment, the self-driving vehicle (SDV) is a self-driving vehicle according to the first aspect of the invention.

[0289] According to a further or alternative embodiment, the self-driving vehicle (SDV) comprises at least one laser scanner, preferably a LiDAR sensor, placed in a fork tip of the load carrier. Preferably, a laser scanner, preferably a LiDAR sensor, is placed in each fork tip of the load carrier, such that the vehicle obtains detailed monitoring of the space behind the forks during backward movement. Preferably, no ultrasonic sensors are present in the fork tips in this embodiment.

[0290] Thus, an independent aspect of the invention also concerns a method for unloading a truck by means of a self-driving vehicle (SDV) comprising the steps of:

[0291] (a) navigating a self-driving vehicle (SDV) up to a truck, preferably to the center of a tailgate of said truck, said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base, wherein the load carrier is loaded with a load, and which truck comprises a cargo space;

[0292] (b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of the cargo space;

[0293] (c) navigating the self-driving vehicle into the loading space of the truck, until the self-driving vehicle has reached a suitable location for loading the load; (d) loading the load from the cargo space of the truck onto the load carrier of the self-driving vehicle; and

[0294] (e) navigating the self-driving vehicle out of the cargo space of the truck, preferably by navigating to the center of the cargo space of the truck, and then navigating away from the cargo space; wherein the navigating and / or maneuvering of the self-driving vehicle is performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising three or more laser sensors and two or more three-dimensional (3D) cameras.

[0295] According to a further or other preferred embodiment, the method uses a system according to the second aspect of the invention.

[0296] According to a further or other embodiment, step (a) is preceded by the step of analyzing the cargo space of the truck. This step is preferably performed by means of the localization camera and / or the plurality of safety sensors. This step allows to check whether loads, preferably pallets, are already present in the cargo space, and if so, where they are positioned. Based on this, the self-driving vehicle can better decide where pallets should be unloaded from the cargo space of the truck.

[0297] According to a further or other embodiment, the method comprises, between steps (a) and (b) and / or between steps (b) and (c), the step of analyzing one or more walls of the cargo space of the truck. This step is preferably performed by means of the localization camera and / or the plurality of safety sensors. This step allows the selfdriving vehicle to orient itself in the cargo space relative to the walls of the cargo space.

[0298] According to a further or other preferred embodiment, the method in step (a) uses the steps as described in the method according to the third aspect of the invention.

[0299] According to a further or other embodiment, step (c) is performed by means of the two or more laser sensors.

[0300] According to a further or alternative embodiment, step (d) comprises the following sub-steps:

[0301] (i) detecting a load to be unloaded from the cargo space of the truck; and (ii) maneuvering the self-driving vehicle, whereby the load carrier is moved under the load.

[0302] Navigation within a cargo space of a truck is particularly challenging as the cargo space is a dark environment, in which light-sensitive sensors such as cameras operate suboptimally. Simultaneous Localization and Mapping (SLAM) can therefore not be optimally used for navigating the self-driving vehicle in the cargo space. Moreover, even if the use of Simultaneous Localization and Mapping (SLAM) were possible, the accuracy of this system is insufficient to allow for the efficient loading of a truck's cargo space. However, the use of the two or more laser sensors does allow unloading of the cargo space.

[0303] According to some embodiments, in step (i) use is made of a pallet camera of the self-driving vehicle (SDV) in order to correctly detect the load to be unloaded.

[0304] According to some embodiments, in step (ii) use is made of force feedback in order to verify whether the load was correctly placed on the load carrier.

[0305] According to a particular embodiment, during unloading, the SDV detects the first object to be unloaded via the pallet camera. The SDV then orients itself relative to this load and drives into the pallet. While this occurs, the LiDAR sensors in the fork tips detect the blocks under the pallet, and count them. If three blocks are determined (i.e., three consecutive field trips), this is considered as full insertion. Subsequently, the load is lifted and brought out.

[0306] Through the combination of LiDAR sensors in the fork tips, field logic, force feedback, and / or configuration-dependent safety strategies, the SDV can operate autonomously, safely, and with high precision in dynamic and narrow environments without the need for additional external control or communication infrastructure.

[0307] According to a further or other embodiment, the self-driving vehicle (SDV) is configured for performing synchronized loading or unloading operations with one or more other SDVs, in particular by means of so-called dual vehicle load sharing. Herein, two SDVs are deployed for jointly handling a load that is heavier or larger than the nominal capacity of a single vehicle, with continuous synchronization of position, speed, and lift height. Furthermore, the SDV is configured for autonomously crossing loading bridges or dock ramps, wherein real-time monitoring of acceleration, inclination angle, and ground stability is used to ensure safe movement between different heights or levels.

[0308] According to a further embodiment of the fourth or fifth aspect of the invention, the method is performed with two self-driving vehicles (SDVs) that simultaneously cooperate in loading or unloading a truck.

[0309] Preferably, the SDVs are coordinated via the multi-agent system (MAS) or a central dispatch module in such a way that they perform complementary tasks within the same loading or unloading process. In the case of loading, this can, for example, entail that both SDVs alternately pick up loads from separate shipping lanes and place them in parallel in the cargo space of the same truck, wherein the path, timing, and positioning of each SDV are coordinated with each other.

[0310] In the case of unloading, the SDVs can preferably sequentially or synchronously remove loads from the truck based on a predefined task allocation. This method allows the loading or unloading time to be considerably shortened and increases operational efficiency, while mutual collisions or conflicts are deterministically avoided based on shared planning and mutual coordination.

[0311] According to some embodiments, the cargo space of the truck offers space in width for three loads positioned next to each other, preferably pallets. More preferably, the method is carried out such that first the middle load, preferably pallet, is unloaded, and thereafter the loads at the side walls are unloaded.

[0312] DESCRIPTION OF THE FIGURES

[0313] Figures 1 to 6 show various views of a self-driving vehicle (SDV) according to an embodiment of the first aspect of the invention, in particular a self-driving vehicle (SDV) comprising: a vehicle base 1, a load carrier 2, which load carrier is coupled to said vehicle base 1. The self-driving vehicle further comprises an industrial computer unit, a programmable logic controller (PLC), a localization camera 3, and a plurality of safety sensors. The industrial computer unit, programmable logic controller (PLC), localization camera 3, and safety sensors are coupled to the vehicle base 1, wherein the industrial computer unit is electronically coupled to the programmable logic controller (PLC), and wherein said localization camera 3 and said safety sensors are electronically coupled to the industrial computer unit and / or the programmable logic controller (PLC). In particular, the safety sensors comprise two laser sensors 4, two ultrasonic sensors 5, and two three-dimensional (3D) cameras 6. The various safety sensors as illustrated are preferably configured for monitoring a monitoring field that is at least 10% larger than a safety field around the self-driving vehicle. Possibly, two laser scanners are located at the fork tips instead of the two illustrated ultrasonic sensors 5.

[0314] Figures 1 to 6 show that said laser sensors 4 are coupled to the front side 7 and / or to the lateral sides 8 of the vehicle, that said ultrasonic sensors 5 are coupled to the rear side 9 of the vehicle, and that said three-dimensional (3D) cameras 6 are coupled to the front side 7 of the vehicle. In particular, it is visible that the vehicle base 1 has a substantially rectangular horizontal cross-section, wherein the laser sensors 4 are located at the two corner points on the front side of the vehicle base 1. Said laser scanners 4 together have a two-dimensional (2D) field of view, which field of view extends in a plane substantially parallel to a ground surface, at a distance comprised between 5 and 20 cm from said ground surface. The vehicle base 1 of the self-driving vehicle (SDV) is further provided with one or more cut-outs 10 at the level of said laser scanners 4. In particular, these cut-outs 10 ensure that said laser scanners 4 together have a two-dimensional (2D) field of view that extends over an angle comprised between 220° and 280° around the self-driving vehicle. The load carrier

[0315] 2 further comprises two forks 12, wherein the ultrasonic sensors 5 are located in the fork tips 11 of the forks 12. Said ultrasonic sensors 5 have a detection range in the form of a three-dimensional (3D) sound beam, which points rearward in the zone behind the fork tips 11 of the forks 12 of the load carrier 1. The localization camera

[0316] 3 as shown herein is positioned on a frame 13 that extends in height above the vehicle base 1, and the localization camera 3 is thereby directed towards the front side 7 of the self-driving vehicle. On this frame 13, a "blue light" 14 is also placed, which "blue light" 14 is directed towards the front side 7 of the self-driving vehicle. This "blue light" 14 contributes to safety as it provides an indication of the direction of travel to persons in the vicinity of the self-driving vehicle. The self-driving vehicle further comprises a pallet camera 15, which is directed towards the rear side 9 of the self-driving vehicle, and which allows a pallet in the vicinity of the self-driving vehicle to be detected. As shown herein, the pallet camera 15 is positioned on the rear side of the vehicle base 1. The three-dimensional (3D) cameras 6 as shown, are positioned on the front side of the vehicle base 1, in particular between the laser scanners 4. Possibly, two laser scanners are located at the fork tips instead of the two illustrated ultrasonic sensors 5.

[0317] Also shown in Figures 1 to 6 are the wheels 16, which are located on the underside of the vehicle base 1, as well as on the underside of the forks 12. The self-driving vehicle also comprises on its vehicle base 1 a number of control elements 17 which allow manual intervention in the control of the self-driving vehicle. An example of such control elements 17 is an emergency stop button. On the vehicle base 1 one or more, in this case two, warning lights 18 are also provided, which can signal certain errors in the control of the self-driving vehicle.

[0318] The illustrated self-driving vehicle is particularly suitable for carrying out a method according to the invention. The method has the advantage that the combination of sensors allows the self-driving vehicle (SDV) to efficiently and controllably navigate and / or maneuver both into and out of the cargo space of the truck. The combination of sensors as described herein further allows obstacles located in the vicinity of the self-driving vehicle to be quickly and efficiently detected. The combination of different types of sensors provides redundancy and robustness in obstacle detection.

[0319] List of elements indicated herein:

[0320] 1 vehicle base

[0321] 2 load carrier

[0322] 3 localization camera

[0323] 4 laser sensor

[0324] 5 ultrasonic sensor

[0325] 6 three-dimensional (3D) camera

[0326] 7 front side of the SDV 8 lateral side of the SDV

[0327] 9 rear side of the SDV

[0328] 10 cut-out of the vehicle base

[0329] 11 fork tip of the fork

[0330] 12 fork of the load carrier

[0331] 13 frame

[0332] 14 blue light

[0333] 15 pallet camera

[0334] 16 wheel

[0335] 17 control elements

[0336] 18 warning light

Claims

CLAIMS1. A method for loading a truck by means of a self-driving vehicle (SDV), comprising the steps of:(a) navigating a self-driving vehicle (SDV) up to a truck, said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base, wherein the load carrier is loaded with a load, and which truck comprises a cargo space;(b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of the cargo space;(c) navigating the self-driving vehicle into the cargo space of the truck, until the self-driving vehicle has reached a suitable location for unloading the load;(d) unloading the load from the load carrier of the self-driving vehicle into the cargo space of the truck; and(e) navigating the self-driving vehicle out of the cargo space of the truck; characterized in that the navigating and / or maneuvering of the self-driving vehicle is performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising three or more laser sensors and two or more three-dimensional (3D) cameras.

2. The method according to claim 1, characterized in that step (a) is preceded by the step of analyzing the cargo space of the truck.

3. The method according to claim 1 or 2, characterized in that the method, between steps (a) and (b) and / or between steps (b) and (c), comprises the step of analyzing one or more walls of the cargo space of the truck.

4. The method according to any of the preceding claims 1-3, characterized in that step (c) is performed by means of the two or more laser sensors.

5. The method according to any of the preceding claims 1-4, characterized in that step (c) comprises the following sub-steps:(i) navigating the self-driving vehicle in the cargo space towards a side wall of the cargo space;(ii) tilting the self-driving vehicle, wherein a corner of the load is moved up against the side wall of the cargo space; and(iii) maneuvering the self-driving vehicle until the load is positioned against the rear wall of the cargo space, or a load behind it, and / or until the selfdriving vehicle and the load are oriented substantially parallel in their longitudinal direction to the side wall of the cargo space; characterized in that during step (ii), the two or more laser sensors monitor the distance between the self-driving vehicle and / or the load, and the side wall of the cargo space.

6. The method according to claim 5, characterized in that step (ii) comprises detecting one or more obstacles around the self-driving vehicle, preferably just before the corner of the load is moved up against the side wall of the cargo space.

7. The method according to claim 5 or 6, characterized in that step (ii) comprises monitoring the distance between the self-driving vehicle and / or the load, and the opposite side wall of the cargo space.

8. The method according to any of the preceding claims 5-7, characterized in that sub-step (iii) comprises scraping and / or dragging the load against the side wall of the cargo space during the maneuvering of the self-driving vehicle, until the load is positioned against the rear wall of the cargo space, or a load behind it, and / or until the self-driving vehicle and the load are oriented substantially parallel in their longitudinal direction to the side wall of the cargo space.

9. The method according to any of the preceding claims 5-8, characterized in that steps (ii) and / or (iii) are performed by means of force feedback from a drive system of the self-driving vehicle.

10. A method for unloading a truck by means of a self-driving vehicle (SDV), comprising the steps of:(a) navigating a self-driving vehicle (SDV) up to a truck, preferably to the center of a tailgate of said truck, said self-driving vehicle comprising a vehicle base and a load carrier, which load carrier is coupled to said vehicle base, wherein the load carrier is loaded with a load, and which truck comprises a cargo space;(b) maneuvering the self-driving vehicle, wherein the load carrier is directed towards the inside of the cargo space;(c) navigating the self-driving vehicle into the loading space of the truck, until the self-driving vehicle has reached a suitable location for loading the load;(d) loading the load from the cargo space of the truck onto the load carrier of the self-driving vehicle; and(e) navigating the self-driving vehicle out of the cargo space of the truck, preferably by navigating to the center of the cargo space of the truck, and then navigating away from the cargo space; wherein the navigating and / or maneuvering of the self-driving vehicle is performed by means of a localization camera and a plurality of safety sensors provided on the self-driving vehicle, said safety sensors comprising three or more laser sensors and two or more three-dimensional (3D) cameras.

11. The method according to claim 10, characterized in that step (a) is preceded by the step of analyzing the cargo space of the truck.

12. The method according to claim 10 or 11, characterized in that the method, between steps (a) and (b) and / or between steps (b) and (c), comprises the step of analyzing one or more walls of the cargo space of the truck.

13. The method according to any of the preceding claims 10-12, characterized in that step (c) is performed by means of the two or more laser sensors.

14. A self-driving vehicle (SDV) for carrying out the method according to any of claims 1-9 or the method according to any of claims 10-13, comprising: a vehicle base, a load carrier, which load carrier is coupled to said vehicle base, an industrial computer unit, a programmable logic controller (PLC), a localization camera, and a plurality of safety sensors, which industrial computer unit, programmable logic controller (PLC), localization camera, and safety sensors are coupled to the vehicle base, wherein the industrial computer unit is electronically coupled to the programmable logic controller (PLC), and wherein said localization camera and said safety sensors are electronically coupled to the industrial computer unit and / or the programmable logic controller (PLC),characterized in that said safety sensors comprise three or more laser sensors and two or more three-dimensional (3D) cameras.

15. A system for controlling a self-driving vehicle (SDV) comprising: at least one self-driving vehicle (SDV) according to claim 14, and a software architecture comprising a user interface (UI), a multi-agent system (MAS), a warehouse management system (WMS), and a warehouse control system (WCS), characterized in that the self-driving vehicle comprises a localization camera and a plurality of safety sensors, wherein said safety sensors comprise three or more laser sensors and two or more three-dimensional (3D) cameras.

Images