Reconfigurable mobile chassis of cabin inspection operation robot based on multi-modal perception
By combining multimodal perception with a reconfigurable chassis, the problems of robot structural expansion and insufficient perception in the cabin environment are solved, enabling efficient and safe autonomous inspection and maintenance tasks.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Utility models(China)
- Current Assignee / Owner
- WUHAN UNIV OF TECH
- Filing Date
- 2025-09-25
- Publication Date
- 2026-08-04
AI Technical Summary
Existing mobile robots in ship cabin environments have fixed structures that are difficult to expand, limited perception methods, unstable path planning, and poor environmental adaptability, resulting in low operational efficiency and high safety risks.
Employing a multimodal sensing device, a chassis inspection motion device, and a reconfigurable chassis device, combined with sensors such as LiDAR, depth camera, and inertial measurement unit, the robot achieves perception fusion and structural reconstruction through the ROS control system, ensuring autonomous inspection in complex environments.
It achieves high-precision environmental mapping, autonomous path planning, and stable motion, improving the robot's adaptability and stability in the ship cabin environment and reducing labor intensity and safety risks.
Smart Images
Figure CN224595032U_ABST
Abstract
Description
Technical Field
[0001] This utility model relates to the fields of mobile robots, intelligent ship cabin operation and maintenance, and perception fusion technology. More specifically, it relates to a reconfigurable mobile chassis for a ship cabin inspection and maintenance robot based on multimodal perception. Background Technology
[0002] With the development of industrial intelligence, the density of equipment in ship cabin environments is constantly increasing, significantly raising the complexity of inspection and maintenance tasks. Currently, inspection work in ship cabins largely relies on manual labor, which is not only labor-intensive and inefficient but also poses safety risks such as high temperature, high humidity, confined spaces, and flammable gas confinement, making the working conditions extremely harsh. Although some existing mobile robots have been applied to simple indoor navigation tasks, they still generally suffer from problems such as fixed and difficult-to-expand structures, limited perception methods, unstable path planning, and poor environmental adaptability in typical industrial environments such as ship cabins. On the one hand, the chassis structure of traditional robots is not adjustable, making it difficult to adapt to narrow passages, variable cabin shapes, and temporary obstacle layouts; on the other hand, perception systems often rely on single lidar or vision sensors, failing to balance planar accuracy and spatial detail, resulting in low mapping quality and frequent navigation failures. Simultaneously, robots face problems such as wheel slippage and idle spinning in the complex metallic environment of ship cabins, making it difficult to correct odometer errors, further reducing path planning accuracy and system stability. Therefore, there is an urgent need for a chassis system for a shipboard inspection and maintenance robot with multimodal perception fusion, motion compensation mechanism and structural reconstruction characteristics, so as to adapt to autonomous inspection and maintenance tasks in complex environments. Summary of the Invention
[0003] The technical problem to be solved by this utility model is to provide a reconfigurable mobile chassis for a ship cabin inspection and maintenance robot based on multimodal perception. Its overall structure is compact, highly reconfigurable, has high perception accuracy, good motion control robustness, and has good engineering deployability and practical application value.
[0004] The technical solution adopted by this utility model to solve its technical problem is: to construct a reconfigurable mobile chassis for a ship cabin inspection and maintenance robot based on multimodal perception, including a multimodal perception device, a chassis inspection motion device and a reconfigurable chassis device. The multimodal sensing device is used to achieve comprehensive mapping of the cabin environment, obstacle detection, and autonomous localization. The chassis inspection motion device is used to ensure that the robot reduces bumps during movement and operation, and realizes autonomous inspection operation; The reconfigurable chassis device is used to enable rapid structural adjustments, functional expansion, and environmental adaptation of the robot according to cabin space and mission requirements.
[0005] According to the above scheme, the reconfigurable chassis device includes an L-shaped connecting plate, a T-shaped connecting plate, inner corner pieces, outer corner pieces, and aluminum profiles; The aluminum profiles form the main body of the chassis frame. The four corners and stress-bearing parts of the main body of the chassis frame are fixedly connected by L-shaped connecting plates and inner corner pieces. The intersection of the crossbeams and longitudinal beams of the main body of the chassis frame is reinforced by T-shaped connecting plates. The high-load area on the outer side of the main body of the chassis frame is structurally reinforced by outer corner pieces.
[0006] According to the above scheme, the chassis inspection motion device includes a hub motor, a driver, casters, a battery, a shock absorber bracket, and a guide rail; The universal wheels are located at both ends of the lower part of the L-shaped connecting plate; the hub motors are located on the lower part of the longitudinal beams on both sides of the chassis frame body and are fixed to the chassis frame body by the shock-absorbing brackets on the side of the hub motors; the battery is located in the middle area of the chassis frame body and is supported and fixed by upper and lower double-layer aluminum profiles; the driver is located at the front crossbeam of the chassis frame body. The guide rail is located in the middle of the upper frame of the chassis frame body, and a power switch and a hub are installed on the guide rail.
[0007] According to the above scheme, the omnidirectional wheel includes two symmetrically arranged at both ends of the lower part of the L-shaped connecting plate; the hub motor includes two symmetrically arranged at the lower part of the longitudinal beams on both sides of the chassis frame body. According to the above scheme, the multimodal sensing device includes a lidar, a radar bracket, a depth camera, an adjustment bracket, an inertial measurement unit, a main control unit, and a hub motor built-in encoder; The lidar is mounted on a radar bracket, and the depth camera is mounted on an adjustable bracket. Both the radar bracket and the adjustment device are fixedly mounted at the center of the front of the chassis frame body to acquire two-dimensional contour information of the cabin. The inertial measurement unit is located in the geometric center area of the chassis frame body and is used to provide attitude change and motion compensation information; The ultrasonic sensors are installed at the four corners of the upper layer of the chassis frame body to achieve near-range obstacle detection and collision avoidance. The main control unit is located in the open area at the front of the chassis frame body and is used to run the control system and perform multimodal data processing. The hub motor has a built-in encoder connected to the driver to acquire chassis speed and displacement information in real time.
[0008] According to the above scheme, the lidar acquires two-dimensional contour information of the environment through 360° planar scanning; the depth camera adjusts the perception angle according to the actual cabin scene, captures three-dimensional images in front, and generates equivalent two-dimensional laser point cloud data through image conversion algorithm to supplement the lidar's perception blind spots on cabin slopes, gaps, and steps.
[0009] According to the above scheme, it also includes a ROS control device, which is used for unified scheduling and collaborative management of perception fusion, mapping and positioning, path planning and motion control in the cabin environment.
[0010] According to the above scheme, the battery is a lithium iron phosphate battery.
[0011] The reconfigurable mobile chassis of the ship's cabin inspection and maintenance robot based on multimodal perception, which implements this utility model, has the following beneficial effects: 1. This utility model adopts a reconfigurable chassis structure design. The whole is built with industrial aluminum profiles. The connectors are diverse and the assembly is flexible. It is easy to adjust the size, sensor layout and function expansion according to different cabin layouts, which improves the structural adaptability, platform versatility and adaptability to inspection and maintenance tasks for the cabin environment. 2. This utility model is designed for industrial applications in ship cabins. The chassis is compact, and the hub motor combined with the universal wheels has high torque turning capability. The large-capacity battery provides stable power supply. The whole system can operate stably for a long time under extreme conditions such as high temperature, high humidity and confined space. It has good engineering practicality and deployment value. 3. This utility model is based on the ROS control system to build a full-process control architecture, integrating SLAM mapping, path planning, motion control and other functions, supporting global path generation and local dynamic obstacle avoidance in a closed cabin, and autonomously adapting to high-precision mobile inspection and maintenance tasks in narrow, winding and multi-obstacle environments. 4. The operation process of this utility model is highly automated, with a closed sensing and control link, accurate positioning, reliable path, compact structure, and flexible response. It is particularly suitable for industrial-grade inspection and maintenance scenarios in ship cabin environments with typical characteristics such as dense equipment, dynamic obstacles, and uneven lighting. Through the organic combination of multimodal sensing fusion and reconfigurable design, the chassis not only achieves autonomous movement in complex spaces, but also significantly enhances environmental adaptability and long-term stability, and has good practical application value and engineering deployment prospects. Attached Figure Description
[0012] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a schematic diagram of the reconfigurable mobile chassis of the ship cabin inspection and maintenance robot based on multimodal perception of this utility model; Figure 2This is a front view of the reconfigurable mobile chassis of the ship compartment inspection and maintenance robot based on multimodal perception of this utility model; Figure 3 This is a side view of the reconfigurable mobile chassis of the ship cabin inspection and maintenance robot based on multimodal perception of this utility model. Figure 4 This is a top view of the reconfigurable mobile chassis of the ship compartment inspection and maintenance robot based on multimodal perception of this utility model; Figure 5 This is a bottom view of the reconfigurable mobile chassis of the ship cabin inspection and maintenance robot based on multimodal perception of this utility model; Figure 6 This is a flowchart of the fusion process of the multimodal sensing module of this utility model; In the diagram: 1. LiDAR, 2. LiDAR bracket, 3. Hub, 4. Power switch, 5. Ultrasonic sensor, 6. Guide rail, 7. Driver, 8. Inertial Measurement Unit (IMU), 9. Battery, 10. Outer corner bracket, 11. Shock absorber bracket, 12. Hub motor, 13. Depth camera, 14. Adjustment bracket, 15. Main controller, 16. Aluminum profile, 17. Caster wheel. Detailed Implementation
[0013] To provide a clearer understanding of the technical features, objectives, and effects of this utility model, the specific embodiments of this utility model will now be described in detail with reference to the accompanying drawings.
[0014] like Figure 1-6 As shown, the utility model discloses a reconfigurable mobile chassis for a ship cabin inspection and maintenance robot based on multimodal perception. This chassis includes a multimodal perception device, a chassis inspection motion device, a ROS control device, and a reconfigurable chassis assembly. A reconfigurable mobile chassis platform with strong perception and structural reconfiguration capabilities is constructed to address the typical characteristics of complex ship cabin structures, limited space, and dense obstacles. A modular assembly system is built using high-strength aluminum profiles 16, allowing for rapid size adjustments and environmental adaptation to different cabin environments. Combined with high-torque hub motors 12 and multi-point shock-absorbing suspension support, the robot maintains high operational stability under typical ship cabin conditions such as swaying, elevation differences, and vibration impacts. Its perception capabilities, structural design, and mapping algorithms are better suited to the precise positioning and stable operation requirements of the complex industrial environment of a ship cabin. Through hardware and software co-design, high-precision mapping, autonomous path generation, and dynamic obstacle avoidance control for ship cabin inspection and maintenance tasks are achieved, providing a complete, deployable, and reconfigurable technical solution for intelligent maintenance in complex ship cabin environments.
[0015] This utility model discloses a reconfigurable mobile chassis for a ship cabin inspection and maintenance robot based on multimodal perception. The chassis includes a multimodal perception device, a chassis inspection motion device, a ROS control device, and a reconfigurable chassis. The multimodal perception device enables comprehensive mapping of the ship cabin environment, obstacle detection, and autonomous localization. The chassis inspection motion device ensures the robot mitigates turbulence during movement and operation, enabling autonomous inspection. The ROS control device provides unified scheduling and collaborative management of perception fusion, mapping and localization, path planning, and motion control within the ship cabin environment. The reconfigurable chassis allows for rapid structural adjustments, functional expansion, and environmental adaptation of the robot based on cabin space and task requirements.
[0016] The reconfigurable chassis assembly includes an L-shaped connecting plate, a T-shaped connecting plate, inner corner pieces, outer corner pieces 10, and an aluminum profile 16. The aluminum profile 16 forms the main body of the chassis frame. The four corners and stress-bearing parts of the main body of the chassis frame are fixedly connected by the L-shaped connecting plate and the inner corner pieces. The intersection of the crossbeams and longitudinal beams of the main body of the chassis frame is reinforced by the T-shaped connecting plate. The high-load area on the outer side of the main body of the chassis frame is structurally reinforced by the outer corner pieces 10.
[0017] The reconfigurable chassis structure is a mechanical support platform specifically designed for environments with limited space, narrow passages, and dense machinery, such as ship cabins. Its core component is based on high-strength industrial aluminum profile 16, assembled into an overall frame through modular connectors to achieve structural adaptability, rapid assembly, and high scalability. The chassis frame employs multiple connection methods, allowing various sensor brackets, battery modules, and wiring devices to be directly fixed to the profile structure. The installation methods are compatible with standard fastening components such as hexagonal spring nuts and Phillips head plates, avoiding the use of welding or irreversible installation methods, thus improving the overall maintainability and reconfigurability of the chassis. To address common vibration and impact issues during shipboard operations, independent shock-absorbing brackets 11 are installed at key load-bearing components of the chassis, such as the mounting location of the hub motor 12. These brackets utilize a combination of a rubber vibration isolation layer and a metal frame, effectively mitigating vibrations caused by motor operation and ground undulations. In this embodiment, the battery 9 is a lithium iron phosphate battery, symmetrically positioned at the center of the chassis. It is suspended below the main frame using double-layer aluminum profiles 16, and high-load support and power stability are ensured through multi-point fastening connections. The battery area also features isolated wiring for power and data lines via nylon cable ties to prevent signal interference.
[0018] The chassis inspection motion device includes a hub motor 12, a driver 7, casters 17, a battery 9, shock absorber brackets 11, and a guide rail 6. The casters 17 are arranged symmetrically at both ends of the lower part of the L-shaped connecting plate. The hub motor 12 is arranged symmetrically on the lower part of the longitudinal beams on both sides of the chassis frame body, and is fixed to the chassis frame body by the shock absorber brackets 11 on the sides of the hub motor 12. The shock absorber brackets 11 are located on the sides of the hub motor 12. The battery 9 is located in the middle area of the chassis frame body, supported and fixed by upper and lower double-layer aluminum profiles. The driver 7 is located at the front crossbeam of the chassis frame body. The guide rail 6 is located in the middle of the upper frame of the chassis frame body, and a power switch 4 and a hub 3 are installed on the guide rail 6. The shock absorber structure deployed on the casters 17 on both sides of the front of the chassis forms a four-point support with the main drive wheel, which enhances ground contact stability and provides lateral sliding and turning capabilities in confined spaces. The chassis features a composite layout with upper and lower layers and a central interlocking structure. The upper layer houses sensors, control systems, and batteries, while the lower layer contains drive wheels, casters, a shock absorption system, and cable channels. This design ensures excellent terrain adaptability for the robot in complex cabin paths, including longitudinal movement, turning on the spot, and lateral fine-tuning. To accommodate different cabin sections or various inspection and maintenance tasks, the chassis is designed with multiple sets of sliding rail holes and fixing slots, supporting lateral and longitudinal profile length adjustment. This allows users to adjust the chassis wheelbase, length, and height according to the usage scenario. Furthermore, the entire chassis platform can be quickly reconfigured into cross-shaped, H-shaped, and other structural layouts by replacing different connectors. This reconfigurable chassis structure is a reconfigurable, scalable, maintainable, and replaceable chassis architecture, greatly improving the robot's adaptability, stable operation, and on-site deployment efficiency in enclosed industrial environments such as ship cabins.
[0019] The multimodal sensing device includes a lidar 11, a lidar bracket 2, a depth camera 13, an adjustment bracket 14, an inertial measurement unit 8, a main control unit 15, and a hub motor 12 with an integrated encoder. The lidar bracket 2 and the adjustment device are fixedly positioned at the center of the foremost part of the chassis to acquire two-dimensional contour information of the cabin. The inertial measurement unit 8 is located at the geometric center of the chassis frame to provide attitude change and motion compensation information. Ultrasonic sensors 5 are located at the four corners of the upper layer of the chassis for near-range obstacle detection and collision avoidance. The main control unit 15 is located in the unused area at the front of the chassis to run the ROS control system and perform multimodal data processing. The hub motor 12 with its integrated encoder is connected to the driver to acquire real-time information such as chassis rotation speed and displacement. The lidar 11 acquires two-dimensional contour information of the environment through 360° planar scanning. The depth camera 13 adjusts the perception angle according to the actual cabin scene, captures three-dimensional images of the front, and generates equivalent two-dimensional laser point cloud data through image conversion algorithms to supplement the lidar 1's blind spots in sensing cabin slopes, gaps, and steps.
[0020] The image conversion algorithm employs the Extended Kalman Filter (EKF) algorithm. The EKF algorithm comprises two parts: prediction and update, corresponding to the motion model and observation model of the mobile chassis. It converts the visual data from the depth camera 13 and the data from the lidar 1 into a locally linear system for fusion. At each time step, a first-order Taylor expansion is used to linearize the system, thereby estimating the robot's current pose and predicting its future state. This improves the robot's accuracy in mapping and localization within the ship's cabin environment. The model expression is: (1) In the formula, For the motion model of the system, it is represented as The state variable at any given time; These are the robot's equations of motion; These are the control inputs, which are represented as linear velocity and angular velocity in this model. This is process noise; (2) In the formula, For the observation model of the system, it represents The observed variables at time; This is a mapping function from observed values to system state variables; The observed noise follows a Gaussian distribution with a mean of 0; For motion model Observation model Perform the corresponding first-order Taylor expansion; (3) (4) In the formula, express The posterior estimate of the state vector at time step; The Jacobian matrix representing the motion model is a function. exist , Regarding The derivative; The Jacobian matrix representing the observation model is a function. exist Regarding The derivative; The motion model of the moving chassis Set as Control input quantity Set as Construct a linear motion model ; (5) In the formula, Indicates the first In the time step, based on the The posterior state estimate of the step is the prediction of the current covariance; In the observation model Below, the observation settings of lidar 1 are... The observation settings of depth camera 13 To obtain the motion model To the observation model Transformation: (6) In the formula, The covariance matrix represents the measurement error of the sensor, i.e., the Gaussian noise. It is a 3×3 identity matrix.
[0021] In the prediction phase, the system state vector for the next time step is predicted using equation (6). The Kalman gain matrix is then calculated using the variance of the sensor measurements and the predicted system values from the update phase. The system's optimal estimate is obtained by combining the current measurement values, and the fused global information is updated.
[0022] The ROS control unit is integrated into the main controller 15, and internally deploys SLAM mapping and path planning algorithms. The SLAM mapping algorithm uses mainstream open-source algorithms such as Gmapping or Cartographer, constructing a 2D map from the generated LiDAR 1 and odometer data. The multimodal perception module at the control layer jointly processes data from LiDAR 11, depth camera 13, IMU8, and encoder, performing dynamic pose estimation via EKF and outputting high-precision odometer-map matching results. All control logic is controlled through ROS nodes using a publish-subscribe mechanism, ensuring timely data sharing and consistent operation among the algorithms, supporting closed-loop information flow in the inspection and maintenance workflow.
[0023] The operation process of this utility model is as follows: After the ROS control device is started in the main controller 15, the initialization of core nodes such as multimodal perception fusion, SLAM mapping, and path planning is completed first, and communication connections with various sensors and drivers 7 are established to ensure the coordinated operation of each module. Subsequently, the multimodal perception device begins to continuously collect environmental data, including planar point cloud information acquired by the lidar 11, three-dimensional image information acquired by the depth camera 13, attitude change data provided by the IMU8, and rotational speed and displacement information fed back by the encoder built into the hub motor 12.
[0024] The aforementioned sensing data is first processed through time synchronization and coordinate registration. Then, multi-source information is fused within the Extended Kalman Filter (EKF) algorithm framework integrated into the ROS system to achieve dynamic estimation of the robot's current pose. The fused localization result is used in the SLAM mapping algorithm to generate an accurate two-dimensional grid map. This map realistically reflects the environmental features of the ship's cabin, including passageways, equipment, and obstacles, providing reliable support for inspection and maintenance operations.
[0025] Once the map is built, the ROS system invokes the Move_base path planning framework to automatically generate a global path based on the robot's current localization and the target task point. The path generation mechanism is specifically optimized for the typical characteristics of the ship's cabin environment, including narrow passages and irregular structures. The robot constructs a high-precision two-dimensional grid map using multi-source data from LiDAR 11, depth camera 13, inertial measurement unit 8 (IMU), and wheel speed encoder. This map accurately reflects the structural layout, passage distribution, and static obstacles within the cabin, providing an environmental basis for path planning. Upon receiving the task objective, the system invokes the global planner to search for the optimal path on this map, avoiding enclosed areas and obstacles, and generating a global reference path from the current position to the target point.
[0026] During robot movement, to cope with sudden changes such as personnel movement and equipment movement within the cabin, the system relies on LiDAR 11 and ultrasonic sensors 5 to perceive the environment ahead in real time. It dynamically adjusts its trajectory using a local planner (such as Teb or DWA), generating local paths in real time to avoid obstacles. The system constructs and continuously updates a cost map to ensure sufficient buffer distance for narrow passage edges and obstacles, preventing collisions. Simultaneously, the system has failure detection and path replanning capabilities during path execution. If a local path becomes impassable, it will automatically attempt fine-tuning, in-situ rotation, or path replanning to ensure stable robot movement in the space-constrained and structurally complex cabin environment, enabling continuous inspection and maintenance tasks.
[0027] The chassis inspection motion device drives the hub motor 12 under the control of path planning commands. Combined with Hall encoders to monitor wheel speed in real time, and based on the PID closed-loop adjustment algorithm, it precisely adjusts the control output to ensure that the robot moves smoothly along the predetermined trajectory. The chassis structure is equipped with omnidirectional wheels 17 to provide lateral support and posture balance in complex path scenarios such as turning and U-turns, enhancing the chassis's mobility and adaptability to local environments.
[0028] During inspection and maintenance operations, all communication on the mobile chassis is achieved via RS485, Ethernet, or USB serial port. Sensing data, navigation status, and execution commands are all published and subscribed to through the ROS system's internal topic mechanism, forming a complete closed-loop process of perception-decision-control. All equipment power is centrally controlled via power switch 4 and powered by battery 9, ensuring continuous and stable operation of the chassis for at least three hours. After the operation is completed, the system automatically plans the return route or continues to execute the next inspection and maintenance task. The entire process is characterized by high automation, strong stability, accurate perception, and good adaptability, making it particularly suitable for unmanned intelligent inspection and maintenance tasks in complex and enclosed spaces such as ship cabins.
[0029] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.
Claims
1. A reconfigurable mobile chassis for a ship's cabin inspection and maintenance robot based on multimodal perception, characterized in that, This includes multimodal sensing devices, chassis inspection motion devices, and reconfigurable chassis devices; The multimodal sensing device is used to achieve comprehensive mapping of the cabin environment, obstacle detection, and autonomous localization. The chassis inspection motion device is used to ensure that the robot reduces bumps during movement and operation, and realizes autonomous inspection operation; The reconfigurable chassis device is used to enable rapid structural adjustment, functional expansion, and environmental adaptation of the robot according to the cabin space and mission requirements. The reconfigurable chassis assembly includes an L-shaped connecting plate, a T-shaped connecting plate, inner corner pieces, outer corner pieces, and aluminum profiles; The aluminum profiles form the main body of the chassis frame. The four corners and stress-bearing parts of the main body of the chassis frame are fixedly connected by L-shaped connecting plates and inner corner pieces. The intersection of the crossbeams and longitudinal beams of the main body of the chassis frame is reinforced by T-shaped connecting plates. The high-load area on the outer side of the main body of the chassis frame is structurally reinforced by outer corner pieces. The chassis inspection motion device includes a hub motor, a driver, casters, a battery, a shock absorber bracket, and a guide rail; The universal wheels are located at both ends of the lower part of the L-shaped connecting plate; the hub motors are located on the lower part of the longitudinal beams on both sides of the chassis frame body and are fixed to the chassis frame body by the shock-absorbing brackets on the side of the hub motors; the battery is located in the middle area of the chassis frame body and is supported and fixed by upper and lower double-layer aluminum profiles; the driver is located at the front crossbeam of the chassis frame body. The guide rail is located in the middle of the upper frame of the chassis frame body, and a power switch and a hub are installed on the guide rail. The multimodal sensing device includes a lidar, a radar bracket, a depth camera, an adjustment bracket, an inertial measurement unit, an ultrasonic sensor, a main control unit, and a hub motor-embedded encoder. The lidar is mounted on a radar bracket, and the depth camera is mounted on an adjustable bracket. Both the radar bracket and the adjustable bracket are fixedly mounted at the center of the front of the chassis frame body to acquire two-dimensional contour information of the cabin. The inertial measurement unit is located in the geometric center area of the chassis frame body and is used to provide attitude change and motion compensation information; The ultrasonic sensors are installed at the four corners of the upper layer of the chassis frame body to achieve near-range obstacle detection and collision avoidance. The main control unit is located in the open area at the front of the chassis frame body and is used to run the control system and perform multimodal data processing. The hub motor has a built-in encoder connected to the driver to acquire chassis speed and displacement information in real time.
2. The reconfigurable mobile chassis of the ship compartment inspection and maintenance robot based on multimodal perception as described in claim 1, characterized in that, The omnidirectional wheels include two symmetrically arranged at both ends of the lower part of the L-shaped connecting plate; the hub motors include two symmetrically arranged at the lower part of the longitudinal beams on both sides of the chassis frame body.
3. The reconfigurable mobile chassis of the ship compartment inspection and maintenance robot based on multimodal perception as described in claim 1, characterized in that, The battery is a lithium iron phosphate battery.