A cleaning robot (and a control method thereof) including a body, a mover, to drive the cleaning robot to move on a ground and perform cleaning work, a detector, configured to detect a ground feature in front of the cleaning robot, a positioner, configured to obtain position information of the cleaning robot, and a controller, electrically connected to the detector and the positioner. The controller stores or obtains a map of a working region of the cleaning robot and can determine a position of the cleaning robot in the map according to the position information of the cleaning robot. Further, the controller recognizes the ground feature according to a detection result of the detector and controls, according to the position of the cleaning robot in the map, the cleaning robot to perform a corresponding action.

In various examples, ultrasonic data augmentations for autonomous and/or semi-autonomous systems and applications are described herein. Systems and methods described herein may use sensor data generated using one or more ultrasonic sensors to generate augmented input data for training one or more machine learning models to generate one or more representations (e.g., one or more maps) of an environment. As described herein, the sensor data may be augmented using one or more techniques such that the augmented input data corresponds to various driving environments (e.g., different driving surfaces), various poses on machines (e.g., different locations and/or orientations), and/or includes additional information associated with the ultrasonic sensor(s) and/or the sensor data. The systems and methods described herein may further use a new architecture to generate input data for the machine learning model(s), where the input data better represents the environment surrounding a machine executing the machine learning model(s).

Techniques for fusing first information generated using one or more learned models with second information generated using one or more non-learned processes to generate third information including one or more updated versions of the first information and/or the second information. In some examples, the first information may indicate one or more locations associated with one or more first objects in an environment, and the second information may indicate one or more attributes associated with one or more second objects in the environment. In some instances, the learned model(s) may generate the first information based at least on first sensor data generated using one or more first sensors of a machine, and the non-learned process(es) may generate the second information based at least on second sensor data generated using one or more second sensors of the machine.

An escaping method for a cleaning robot having a surface medium sensor includes detecting a surface medium change signal in response to the robot encountering an obstacle and turning along an edge of a first surface medium area. The robot searches a pre-established room map to determine whether a second surface medium area exists in the map and performs an escaping strategy. If the second surface medium area is present, the robot determines whether a bypass route exists based on the map and the boundary of the second area. The robot travels along the route if it exists, or returns along a cleaned route if it does not. If the second surface medium area is not in the map, the robot scans and stores edge information. The method also includes detecting entry into the second area and retreating if entry occurs.

Intelligent algorithms and systems (vehicles and/or mechanisms) locate and extract economic sound concentrations of e.g. any combinations of nodules, manganese crusts and/or sulphide deposit, and separate uneconomic matter from valuable minerals by applying differences in electric and/or acoustic properties to differentiate economically valuable minerals from cost bearing unprofitable other matters (e.g. mud, gravel, rocks, organic matter), thus providing added profitability compared to existing mining machines. Complex and multiple sophisticated technological fields, including, but not limited to Geophysics, Advanced sensor technology (acoustic and electric parameter detection), Signal processing (feature extraction and pattern recognition), Machine learning/AI (classification algorithms and adaptive systems), Mechanical engineering (precision collection mechanisms), Real-time control systems (feedback-based operation), Economic modeling (dynamic threshold determination) are combined. Both independent systems and add-on vehicles to existing mining machines have been developed. Environmental impacts are minimized by the nature of the invented technical solutions. MS and/or AI methods and algorithms can be incorporated.

An autonomous system for detecting, localizing, and potentially deactivating chemical threats or emissions using multiple sensing modalities and reinforcement learning techniques. The system includes visual sensors (e.g., RGB, RGBD, LIDAR), non-visual sensors (e.g., gas concentration, airflow, GPS, RADAR), a neural network architecture and processor to fuse information from different sensors, a module based on deep reinforcement learning for decision making, and a robotic interface for executing actions. The neural network extracts relevant information from sensor streams and encodes them into a joint embedding space. The module considers the current observations, historical data, and previous actions to determine the optimal action for threat localization under partially observable conditions. The system is trained in simulated environments to minimize source localization time while accounting for various constraints. The autonomous system enables effective chemical threat detection and source localization in complex, dynamic environments without endangering human operators.

A method for determining the pose of a pallet relative to an industrial truck that has a pair of fork arms with a first and a second fork arm and a first distance measuring means that is arranged in the region of a fork tip on the first fork arm and preferably directed towards a fork gap, wherein the pallet has an outer block, outer web, center block or web,

characterized by the steps of: inserting the industrial truck into the pallet so that the first and second fork arms enter into the pallet, determining multiple first distance measurement values with the first distance measuring means to the outer block, outer web, center block or web of the pallet during insertion, calculating a degree of offset from at least one first distance measurement value and/or a degree of tilt from the difference between at least two first distance measurement values.

A method and a system for ascertaining a surface condition of a road region using at least one sensor of a vehicle. Data from the sensor are collected, wherein the collected data are evaluated in order to ascertain the surface condition of the road region. The surface condition is analyzed with regard to its shape, and at least one road defect of the road portion is ascertained. A planned vehicle trajectory of the vehicle is adjusted based on the at least one ascertained road defect in order to reduce the ascertained road defect and/or to prevent further road defects.

Provided is a crewless aircraft capable of accurately estimating its own position even inside a manhole, as well as a crewless aircraft control method. A crewless aircraft according to the present invention is a crewless aircraft used to inspect an interior of a manhole, and includes: a camera sensor that captures an image of a manhole opening; a plurality of rangefinders that measure a distance to a ground surface or a predetermined surface in the interior; and a control unit that estimates an own position on the basis of recognition information of the manhole opening obtained by performing recognition on image information obtained from the camera sensor, and distance information of the distance to the ground surface or the predetermined surface obtained from the rangefinders.

A mobile security robot includes a human-sized mannequin mounted on a vehicle. A storage unit, mounted on the vehicle, stores security devices and high-powered energy storage devices for facilitating extended patrols without recharge. A video recording system, disposed in the mannequin, continuously records images of a patrol area. Multiple sensors mounted on and proximal to the mannequin generate sensor data based on environmental conditions of the patrol area. A computing system coupled to the sensors processes the sensor data using artificial intelligence models and generates action commands for execution of tasks by actuators including electric motors, robotic arms, and supplementary attachment devices. The electric motors run the vehicle at different speeds with wheel speed feedback based on the environmental conditions and navigate the vehicle along a predefined travel path with object avoidance during patrols. User interface devices facilitate auditory and visual communication with humans in the patrol area.