Device for dynamic product presentation using AI-controlled, actuator-based textile deformation

Abstract

Device for dynamic product presentation, comprising: a) a base plate (10) as a support structure; b) a plurality of linearly movable actuators (12a) arranged in a matrix-shaped arrangement (12) perpendicular to the base plate (10) and each individually controllable; c) a flexible textile cover (14) stretched over all the actuators (12) and resting on the upper ends of the actuators; d) an electronic control unit (20) for individually controlling each actuator; characterized in that the control unit (20) comprises a trained machine learning model, in particular an artificial neural network, which is trained and configured to capture inputs in the form of natural language commands or free text and to translate these semantic descriptions in real time into coordinated control signals for a dynamic, freely definable three-dimensional surface topography of the textile cover (14).

Description

Technical field

[0001] The utility model relates to a device for the visually appealing and dynamic presentation of products, particularly luxury goods such as handbags, jewelry, perfume bottles, watches, or similar high-value items. The device uses a matrix-arranged plurality of linearly movable actuators to create a controlled deformation of a flexible textile surface on which the product to be presented is placed. The complex coordination of the individual actuators' movements is achieved via a central control unit with an AI module that translates natural language or free text into dynamic, freely definable three-dimensional movement patterns. State of the art

[0002] In the field of product presentation and window display design, various methods are known to draw the attention of potential customers to displayed goods. Conventional approaches include static displays, rotating pedestals, or projection-based solutions.

[0003] Pin-based surfaces are known from the field of kinetic art, where a multitude of vertically movable rods form a three-dimensional surface. However, such installations are primarily conceived as independent works of art. Furthermore, textile deformation systems from stage technology are usually based on cable pulleys or pneumatic systems that do not allow for fine-grained, matrix-based control.

[0004] Existing matrix systems for surface deformation are technologically very limited, as they rely on fixed, pre-programmed animation loops or rigid mathematical functions (such as parameterized sine waves) to generate movements. This significantly restricts design freedom and requires programming effort for every adaptation. Problem statement

[0005] The invention is based on the objective of providing a product presentation device that enables high-resolution, dynamic deformation of a textile surface on which a product is presented. The system is expressly not intended to be limited to predefined, purely wave-like or repetitive movement patterns. Rather, an intuitive control system is to be created that is capable of transferring abstract human concepts and complex choreographic movement sequences onto the textile surface without requiring the writing of classical program code or the parameterization of formulas. Solution to the problem

[0006] The task is solved by a device that combines a mechanical actuator matrix with an adaptive, AI-controlled logic level.

[0007] The hardware component comprises a base plate (10) as a support structure on which a plurality of individually controllable, vertically movable actuators (12a) are arranged in a regular grid as an actuator matrix (12). The actuators (12a) are preferably arranged at a grid spacing of 5 mm to 50 mm and have a maximum vertical stroke of 10 mm to 300 mm in order to be able to reproduce high-resolution three-dimensional topographies. A flexible textile surface (14) is stretched over all of these actuators and preferably fixed to the base plate (10) by means of a circumferential tensioning frame (15), which rests on the upper, preferably rounded, ends of the actuators. The presentation material (30) is placed on this textile surface.

[0008] The upper ends of the actuators (12a) are preferably designed as convexly rounded contact heads, the surface of which is polished or coated with a friction-reducing material, for example PTFE. This enables low-friction sliding of the textile support (14) over the contact heads during lateral stretching. The textile support (14) rests loosely, i.e., without local fixation, on the contact heads, thus creating a defined force-fit connection: The pretension of the tensioning frame (15) presses the textile against the actuator ends, while the rounded geometry and the low-friction coating prevent damage to the textile fibers during relative movements. This contact design ensures the mechanical feasibility of high-resolution textile deformation (see Fig. 2, Detail X).

[0009] The core functional solution lies in the control unit (20). This unit incorporates an implemented, trained machine learning model, specifically a neural network, which is designed to capture user requests in the form of natural language commands, free text, or speech prompts. The AI ​​module translates these abstract semantic inputs in real time into coordinated, spatiotemporal movement profiles (trajectories) for each individual actuator in the matrix. The calculated trajectories are directly converted into electrical control signals via dedicated power electronics or a multi-channel motor driver (26). These signals cause the physical vertical movement of the individual actuators (12a) and thus the actual deformation of the textile overlay (14). The AI ​​module is therefore not an isolated software module, but an integral functional component of the physical chain of effects from the semantic input to the mechanical surface deformation.

[0010] Instead of relying on rigid wave equations, the AI ​​autonomously generates stochastic, organic, or thematic movement patterns from the input context. Commands such as "gentle breathing," "restless sea surface," "pulsating heartbeat," or "explosion from the center outwards" are interpreted by the control unit (20) and translated into fluid raising and lowering of the textile overlay. The AI ​​module can be connected either locally on the control unit, via a connected input interface (16) (e.g., a smartphone sending a wireless signal (18)), or via a cloud architecture. Advantages of the invention

[0011] Limitless creative freedom: a departure from rigid wave patterns towards organic, emotionally effective textile movements.

[0012] Intuitive operation: No programming knowledge is required to create new animations. Adaptation to new products or campaigns takes only seconds via voice or text input.

[0013] High-quality aesthetics: The fine-granular control of the matrix in combination with drapable textiles (e.g. silk, velvet) creates a flowing look that underlines the value of high-quality goods. III. List of reference symbols 10 Base plate (support structure) 12 Actuator Matrix 12a Single actuator 14 textile layer (flexible material layer) 15 tension frames for textile fixation 16 Input interface (external control unit) 18 Wireless signal 20 Control unit with integrated AI module 22 connection ports 24 Interaction area 26 Motor drivers (power electronics) 30 presentation items d Grid spacing of the actuators h_max Maximum vertical stroke IV. Summary

[0014] The invention relates to a device for dynamic product presentation ( Fig. 1) The system comprises a base plate (10) with a matrix-shaped arrangement (12) of linearly movable actuators (12a) and a flexible textile surface (14) stretched over it. A presentation item (30) placed on the textile surface (14) is highlighted by the deformation of the fabric. The complex movement coordination is controlled by a central control unit (20) which incorporates a trained machine learning model. This module interprets abstract, natural language commands (e.g., via an input interface (16)) in real time and translates them directly into coordinated control signals for the power electronics (26) of the actuators (12a), thereby generating a physical, three-dimensional surface topography of the textile surface. V. Description of Figures (drawing sheets are attached separately) Fig. Figure 1: Top view of the system. The top view shows the base plate (10) with the matrix-shaped arrangement of the actuators (12). The textile surface (14) is shown with a dashed line because it covers the actuators underneath. The tensioning frame (15) secures the textile surface around its perimeter. The interaction area (24) is marked as a dash-dotted rectangle on the textile surface and defines the area in which the presentation item (30) is placed and the dynamic deformation takes place. Commands are transmitted via a wireless signal (18) through an external input interface (16) to the control unit (20) with integrated motor driver (26). The grid spacing (d) between the actuators is dimensioned. Fig. 2: Sectional view AA of the surface topography. The section illustrates how the actuators (12a), extended to varying degrees, deform the flexible textile support (14). The presentation item (30) rests in a recess calculated by the control unit. The tensioning frames (15) are shown as clamping profiles at the edges. The dimensions for grid spacing (d) and maximum stroke (h_max) are indicated. Detail X shows the contact zone between the rounded actuator head (12a) and the textile support (14): The convex head geometry allows the textile support to slide with low friction under lateral tension, without local fixation. Fig. 3: System overview and signal flow. The block diagram shows the signal flow in five numbered stages: (1) The user enters a natural language command via the input interface (16). (2) The wireless signal (18) transmits the input to the control unit (20), whose AI module performs the semantic interpretation. (3) The calculated trajectories are converted into digital control signals for each individual actuator in the signal processing unit. (4) The motor driver (26) converts the digital signals into electrical power signals. (5) The actuators (12a) of the matrix (12) move individually and generate the physical textile deformation. Optionally, the AI ​​processing can be outsourced to a cloud architecture (shown with dashed lines). Fig.4: Perspective view in action. The isometric view shows the base plate (10) with the tensioning frame (15) and the actuators (12a) of the matrix (12) arranged therein at different height positions. The textile surface (14) forms an organic surface topography on whose crest the presentation material (30) is dynamically presented. Arrows indicate the vertical movement of the actuators. The input interface (16) sends a command to the control unit (20) via a wireless signal (18). The motor driver (26) integrated into the control unit is shown as a separate section and generates the electrical control signals for the physical deformation.

Claims

Device for dynamic product presentation, comprising: a) a base plate (10) as a support structure; b) a plurality of linearly movable actuators (12a) arranged in a matrix-shaped arrangement (12) perpendicular to the base plate (10) and each individually controllable; c) a flexible textile cover (14) stretched over all the actuators (12) and resting on the upper ends of the actuators; d) an electronic control unit (20) for individually controlling each actuator; characterized in that the control unit (20) comprises a trained machine learning model, in particular an artificial neural network, which is trained and configured to capture inputs in the form of natural language commands or free text and to translate these semantic descriptions in real time into coordinated control signals for a dynamic, freely definable three-dimensional surface topography of the textile cover (14). Device according to claim 1, characterized in that the machine learning model comprises a transformer-based motion generator or a diffusion model configured to transform abstract or metaphorical motion descriptions into individual, time-synchronous trajectory profiles of the individual actuators (12a) in order to generate organic or stochastic surface deformations independently of predefined mathematical wave functions. Device according to claim 1 or 2, characterized in that the machine learning model is optionally implemented locally in the control unit (20), on an external input interface (16) or on a remote server structure. Device according to one of the preceding claims, characterized in that the input of the natural language commands or free text is transmitted via a wireless signal (18) from an external input interface (16) to the control unit (20). Device according to one of the preceding claims, characterized in that the textile covering (14) consists of a material with high elasticity and drapability, preferably silk, satin or velvet. Device according to one of the preceding claims, characterized in that the device further comprises a circumferential tensioning frame (15) by which the textile support (14) is fixed over the actuator matrix (12) with a defined pretension. Device according to one of the preceding claims, characterized in that the upper ends of the actuators (12a) are designed as convexly rounded contact heads, the surface of which is polished or coated with a friction-reducing material to ensure low-friction sliding and protection of the textile covering (14). Device according to one of the preceding claims, characterized in that the actuators are designed as linear servomotors, stepper motors with spindle drive, solenoid actuators or piezoelectric actuators and are arranged at a regular grid spacing (d) of 5 mm to 50 mm and have a maximum vertical stroke (h_max) of 10 mm to 300 mm. Device according to one of the preceding claims, characterized in that the control unit (20) has a multi-channel motor driver (26) with power electronics which converts the calculated digital control signals into electrical power signals for individual height adjustment of each individual actuator (12a). Device according to one of the preceding claims, characterized in that the device further comprises a lighting system and / or a camera system which is synchronized by the control unit (20) with the generated movement sequences of the textile surface (14).

Images