45results about How to "Improve anthropomorphism" patented technology

The invention discloses an automatic driving overtaking decision-making method based on reinforcement learning under opposite double lanes. The method comprises the following steps of collecting the traffic state of an automatic driving vehicle through a sensor; inputting the collected traffic state into a trained decision model; enabling the decision model to select a corresponding driving actioninstruction from the action space according to the input information and output the driving action instruction, and driving the vehicle automatically to form a new traffic state after the driving action; calculating a reward value of the driving action through a reward function, and storing the original traffic state, the driving action, the reward value and the new traffic state into an experience playback pool as transfer samples; calculating a loss function value of the decision model, and optimizing parameters of the decision model according to the transfer sample and the loss function value; and repeating the steps until the automatic driving is finished. The safety and comfort of the overtaking decision-making process of the automatic driving vehicle are ensured, and the humanization and robustness of decision-making are improved through a reinforcement learning decision-making method.

The invention relates to a smart under-actuated bionic robot finger device with parallel-connected tendon ropes, which belongs to the technical field of anthropomorphic robots. A diarticular finger device comprises a pedestal, a first motor, a second motor, a juxta-articular shaft, a first finger segment, a distal-articular shaft, an tail end finger segment and a return spring. The device also comprises a first rope wheel, a second rope wheel, a third rope wheel, a first tendon rope, a second tendon rope and a third tendon rope. A polyarticular finger device also comprises at least one middle finger segment and at least one middle rope wheel. The device comprehensively realizes the special effect of combining finger original configuration variability and self-adaptive grasp by utilizing the motors, the rope wheels, the tendon ropes and the return spring. The device can flexibly bend the middle joints of fingers before grasp to achieve a stable and anthropomorphic prebent gesture and grasps an object in a self-adaptive under-actuated mode when in grasp. The grasping action of the device more approaches to a human hand, and the device can self-adaptively and stably grasp different objects and is applicable to an anthropomorphic robot hand.

The present invention provides a method for generating a robot interactive content, comprising acquiring an expression information of a user; acquiring a user's text emotion information; determining a user's intention based on the expression information and the text emotion information; determining, based on the expression information, the text emotion information, The user intends to combine the current robot life timeline to generate robotic interactive content. The invention adds the life timeline of the robot to the interactive content generation of the robot and makes the robot more humanized when it interacts with the human, so that the robot has the human life way in the life timeline, and the method can enhance the robot interaction content generation of human nature, to enhance human-computer interaction experience, improve intelligence.

The embodiment of the invention discloses a method and device for processing information. An embodiment of the method comprises the following steps: generating response voice information based on voice information sent by a user; generating a phoneme sequence according to the response voice information; generating mouth shape change information according to the phoneme sequence, wherein the mouthshape change information is used for controlling the mouth shape change of a displayed three-dimensional portrait when playing the above response voice information; and playing the response voice information, and controlling the mouth shape change of the three-dimensional portrait according to the mouth shape change information. Through the method disclosed by the embodiment of the invention, themouth shape change of the three-dimensional portrait is controlled based on the played response voice information when the playing of the above response voice information is realized, and a personification effect of the three-dimensional portrait is improved.

The present invention provides a method for generating a robot interactive content, comprising: acquiring a multimode signal; determining a user's intention based on the multimode signal; combining the multidimensional signal and the user's intention, in combination with the current robot variable parameters generate robot interactive content. The invention adds the variable parameter of the robot to the interactive content generation of the robot so that the robot can generate the interactive content according to the previous variable parameter, so that the robot can be more anthropomorphic when interacting with the human being so that the robot can live shaft has a human way of life; the method can enhance the robot interactive content generated anthropomorphic, enhance the human-computer interaction experience and improve intelligence.

The embodiment of the invention discloses a method and device for generating information, and relates to the field of cloud computing. A specific embodiment of the method comprises the following steps: receiving a video and an audio of a user sent by a client; analyzing and processing the video and the audio, and determining the intention category of the user; generating feedback information according to the intention category of the user and a preset service information set; based on the feedback information, generating a pre-established video of the three-dimensional virtual portrait throughan animation engine; and sending the video of the three-dimensional virtual portrait to the client, so as to enable the client to present the video to the user. According to the embodiment, the videoof the three-dimensional virtual portrait meeting the intention category of the user is fed back to the client, and the field interaction accuracy of the user and the three-dimensional virtual portrait is improved.

The invention discloses a two-degree-of-freedom humanoid ankle joint. The two-degree-of-freedom humanoid ankle joint comprises a upper shank part, a lower shank part, a driving device I, a driving device II, a double-jointed component, a foot and a brake cable; the double-jointed component comprises an ankle joint driving device and a subtarlar joint driving device; the ankle joint driving device comprises an ankle joint shaft; the subtarlar joint driving device comprises a subtarlar joint shaft; the ankle joint comprises an upper shaft and a lower shaft, wherein the upper shaft is the ankle joint shaft, and the lower shaft is the subtarlar joint shaft; and the ankle joint shaft and the subtarlar joint shaft are in space cross, and the axes of the ankle joint shaft and the subtarlar joint shaft are not in the same plane. The two-degree-of-freedom humanoid ankle joint provided by the invention can be used for a humanoid robot and improves the personification of a humanoid robot. The mechanism can achieve three movement angles by controlling two parameters, namely, the space two-degree-of-freedom mechanism approximately achieves the three-degree-of-freedom movement of the ankle joint and improves the flexibility of the movement.

A game engine device based on HTML5 comprises an operation environment part and a development tool part, the operation environment part comprises a scene manager suitable for processing nodes of a scene contained in to-be-processed data and converting setting of the nodes into rendering instructions, wherein the rendering instructions comprise material information; and an image rendering module which is suitable for combining rendering instructions with the same material information, and rendering by using the combined rendering instructions. Through the scheme provided by the invention, the basic characteristics of HTML5 can be supported, the platform difference is shielded, a game developer can develop a game conveniently, and the game performance and the user experience are improved.

The embodiment of the invention discloses an emotional intelligent voice call-out method and device, a server and a storage medium, and the method comprises the steps: obtaining a customer service corpus and a customer corpus of a first flow node, converting the customer service corpus into first customer service voice data, and transmitting the first customer service voice data to a target customer; receiving voice data of a target client, and determining a target process baseline where the first process node is located according to a matching result of a target client text corresponding to the voice data and the client corpus; determining a next process node according to the process sequence value of the first process node in the target process baseline; obtaining an emotion label of thevoice data according to the voice feature information and the emotion text information of the voice data; obtaining a target emotion label according to the emotion contrast matrix and the emotion label; and generating second customer service voice data according to the target emotion label and the customer service corpus of the next process node, and replying the second customer service voice data to the target customer. By adopting the method and the device, the outbound success rate and the anthropomorphic degree of the outbound voice can be improved.

The invention discloses a collaborative processing method and device for an intelligent robot. The intelligent robot is provided with a robot operating system. The method includes the steps of receiving and analyzing a multi-mode input data and according to an analyzing result of the multi-mode input data to determine whether other intelligent terminals need to work together or not; sending a task request to a corresponding intelligent terminal when other intelligent terminals need to cooperate. A task request carries a request processing task information, combining task processing results of an intelligent device feedback to execute multi-mode output. The intelligent robot is capable of sharing information with a plurality of robots in the process of performing more complicated tasks to realize a cooperative interaction and improve anthropomorphism and intelligence.

The invention discloses a multi-mode output method and apparatus applied to an intelligent robot. The multi-mode output method comprises the steps of receiving multi-mode input information; analyzing the multi-mode input information, and generating voice text information corresponding to the multi-mode input information according to an analyzing result; extracting specific vocabulary from the voice text information, and generating an action instruction matched with the specific vocabulary; and completing voice output and intelligent robot action output according to the voice text information and the action instruction. By adoption of the multi-mode output method and apparatus, the intelligence and personification capability of the robot can be improved, so that interaction experience between a user and the robot can be improved.

The invention provides a method and system for constructing a three-dimensional space based on voxel data, a server and a computer readable storage medium. According to the invention, pixel data is converted into voxel data so as to be highly matched with GPU operation; layer data and connection data are designed according to the voxel data, and a parallel computing speed is higher in virtue of the layer data and the connection data, so compared with an existing CPU operation mode, the performance of a data operation scheme based on GPU is 2-3 orders of magnitude higher, and an AI algorithm is finer and higher in anthropomorphic degree; in a dynamic updating process, due to the fact that the parallel computing speed is high, the data of a global super-large map of a game can be rapidly updated; and the whole set of data is divided into static data with high memory occupation and dynamic data with low memory occupation, so video memory space occupied in the whole game process is small, time for refreshing the dynamic data each time is extremely short, and the game is smoother.

The invention discloses a robot mouth five-rod moving mechanism which comprises a machine frame. Two motors are arranged on the machine frame, output shafts of the motors are connected with a mouth module which is a five-rod moving mechanism body. The mouth module is composed of an upper lip fixing flake, an upper lip left moving flake, an upper lip right moving flake, a lower lip left moving flake and a lower lip left moving flake. The upper lip fixing flake is supported on the output shafts of the two motors through bearings. The output shafts of the two motors are in key connection with the upper lip left moving flake and the upper lip right moving flake respectively. The whole mouth mechanism can move through output of the motors. The lower lip left moving flake and the lower lip right moving flake are hinged to the upper lip left moving flake and the upper lip right moving flake respectively. The five-rod moving mechanism is simple in structure, can well achieve opening and closing motions of the mouth of a robot, and accordingly various mouth shape motions are achieved.

The invention provides a robot multi-dimension response interaction method and device. The method comprises the steps that outside input information of at least one dimension is obtained; the outsideinput information is subjected to data processing; and according to the data processing result and robot preset data and robot historical data, a response data set of a robot is determined, and activeinteraction with a user is achieved according to the response data set. According to the provided robot multi-dimension response interaction method and device, response can be achieved for the user input, active interaction with the user can be achieved in combination with the outside environment changes, activity is higher, and human simulating performance of the robot is improved.

The invention discloses a human arm kinematics modeling method based on Gaussian process learning. The method comprises establishing a seven-degree-of-freedom (DOF) human arm kinematics model, inputting the palm pose of the human arm and the lengths of the upper arm and the lower arm, and outputting the most comfortable arm configuration when the human arm moves. According to the method, an infrared motion capture instrument is used for collecting shoulder joints, elbow joints, wrist joints, palms central pose and the lengths of upper and lower arms when human arms move in a most comfortable manner; a Gaussian process regression method is used for learning a Gaussian process regression model of a palm center pose, upper and lower arm lengths and an arm angle when a human arm moves in a most comfortable mode, and a human arm kinematics model based on Gaussian process learning is established. The method has the advantages that the complex motion mechanism of the human arm does not need to be studied, and compared with traditional methods based on energy optimization, biomechanical models and the like, calculation is simpler; the palm center pose is used as input, the lengths of the upper arm and the lower arm are used as input, and therefore the anthropomorphic arm obtained through the method is more accurate in configuration and higher in universality.

The invention provides a navigation method and system for three-dimensional space based on voxel data, and a computer readable storage medium. According to the invention, pixel data is converted into voxel data, so the voxel data is highly matched with GPU operation; layer data and connection data are designed according to the voxel data, so a parallel calculation speed is higher in virtue of the layer data and the connection data; through the connection data, passing between different voxel layers on different foundation columns and different layers of adjacent foundation columns is achieved, and therefore, movement of an object in the three-dimensional space is achieved; different connection data can support different passing modes, a plurality of passing paths formed by the connection data exist between two fixed starting points and end points, and different connection data support different passing modes, so a plurality of passing combination modes exist between the two fixed starting points and the end points; and by selecting a combination mode with the minimum cost, an optimal passage combination mode can be selected.