Advanced humanoid robot with built-in computer for real-time application

Abstract

A coordinated control system designed to have a small form factor can be located within a robotic or automated online manufacturing system, the coordinated control system comprising: on-board advanced distributed control hardware configured with artificial intelligence enabled processors incorporated within a System-on-Module (SOM) board, each SOM comprising four processors or nodes; wireless communication with cameras, motors, general purpose I / O via I2C extenders, audio interfaces, the Internet, such as Wifi and Bluetooth connections, for transmitting and receiving commands and any other data; a sensor electrically connected to a plurality of sensors for perceiving an environment and collecting data from infrared, haptic, proximity and other types of sensors; the external host flash memory computer is specially used for uploading / downloading firmware, cloning and configuring; the nonvolatile memory is used for storing a control algorithm, a configuration file and other necessary data; the HDMI-based user interface is used for robot operation, training and setting; the internal star network based on USB C or Ethernet is used for bypassing the standard PCI bus interface BUS to carry out high-speed communication; an Ethernet switch enables a plurality of boards to access the Ethernet for internal communication within the star network as well as external Internet access.

Description

Technical Field

[0001] The present invention relates to a system and method for increasing the operating speed of robots employed in real-time applications by implementing an Advanced Distributed Computing Hardware System (ADCH) having a small form factor, the ADCH being capable of performing parallel computations for multitasking. The present invention also provides an implementation of an end-to-end autonomous stand-alone robot by incorporating the ADCH into a robot including a digital signal processor board, the robot being capable of running multiple neural networks in parallel and processing a wide range of data from a vision system, an array of interface devices, and sensors supported by artificial intelligence inference, deep learning, and reinforcement learning software algorithms, thereby resulting in a powerful and efficient system. Background Art

[0002] A robot is an electromechanical assembly controlled by one or more computer programs and / or electronic circuits. An autonomous robot can perform desired tasks in an unstructured environment without continuous human intervention. In contrast, semi-autonomous and non-autonomous robots typically require human intervention in the form of training to load, unload specific objects, or perform activities such as sorting, processing, packaging, etc. thereon. Robots are used in a variety of fields, including, for example, manufacturing, space exploration, pharmaceuticals, surgery, automotive, etc. Special-purpose robots are typically designed to perform a single task or a single set of tasks, such as moving from one point to another and performing multiple tasks along a route or at a given destination before being instructed to perform a new task. A humanoid robot is a type of robot that attempts to emulate certain human tasks, including but not limited to messy and dangerous work among many other tasks. A humanoid robot is typically constructed to have anthropomorphic characteristics capable of understanding and interpreting human commands through movement. A humanoid can be designed for functional purposes, such as interacting with human tools and the environment or performing more intelligent tasks through the use of artificial intelligence and continuous machine learning. In some cases, a humanoid robot may also have a head designed to replicate human sensory features such as eyes or ears, so that they can be programmed to view, perceive, and understand the operating environment, hear instructions, and then act on them based on what they are programmed to do.

[0003] Traditional humanoid robots generally do not have built-in computing capabilities similar to high-end computers or servers. Instead, humanoid robots typically rely on external computing systems to perform complex tasks and computations while they perform simple tasks. Traditional humanoid robots generally include a combination of hardware components, sensors, and actuators, as well as an embedded system that implements basic processing and control functions. These embedded systems are responsible for handling low-level tasks such as motor control, sensor data processing, and actuator control.

[0004] Humanoid robots are forced to use artificial intelligence (AI) to improve their performance to make them more self - sufficient and intelligent. AI - based humanoid robots are designed to resemble and interact with humans. It combines artificial intelligence (AI) technology to enable it to perceive its surroundings, make autonomous decisions, and perform multiple tasks. Effective collision - avoidance strategies for overcoming obstacles usually combine multiple methods, using infrared and proximity sensor arrays, mapping, path planning, and dynamic adjustment to ensure safe, collision - free, and efficient robot movement. Such robots are equipped with various infrared, tactile, proximity, and other types of sensors, such as cameras, microphones, and touch sensors, to gather information from their environment. AI algorithms process such sensory inputs to understand the operating environment and make informed decisions. However, processing environmental information requires fast analysis, deep learning, reinforcement learning, and machine - learning software modules to handle the computationally intensive variety of data. Conventional systems utilize general - purpose microprocessors that usually exist in external servers, which process the information and send the results to the AI robot, resulting in time delays. Due to their inherent hardware configuration, the processing of commands occurs sequentially with very little or no parallel processing. This affects the speed of the robot, which translates into low productivity and efficiency.

[0005] The AI capabilities of humanoid robots can vary widely. Some robots are programmed with predefined behaviors and responses, while others use machine - learning techniques to learn from their experiences and improve their performance over time. Deep - learning algorithms (a type of machine learning) have been used to enhance the cognitive abilities of humanoid robots, enabling them to recognize objects and faces, hear, understand, and generate speech, and even display emotions.

[0006] The applications of AI - based humanoid robots are diverse. They can be used in fields such as healthcare, where they can assist in patient care and rehabilitation exercises. They can also be used in education, where they serve as interactive tutors or companions for children with special needs. In the customer - service field, humanoid robots can provide assistance and information in public spaces such as airports or shopping malls.

[0007] However, when it comes to more advanced computing and decision-making processes, humanoid robots typically rely on external computing resources. These resources can include cloud-based servers or remote computers that handle heavy computational loads. The robot sends sensor data to the external computing system, which analyzes the information, performs complex calculations or analyses, and sends instructions or commands back to the robot. By leveraging external computing capabilities, humanoid robots benefit from a wider range of computational power, access to large amounts of data, and the ability to utilize advanced algorithms and machine learning models. This approach allows for greater flexibility and scalability in the tasks and applications that humanoid robots can perform. However, as humanoid robots are required to use artificial intelligence and deep learning algorithms to perform increasingly intelligent tasks, the amount of data being analyzed starts to increase, and external servers cannot cope with the timing constraints. There will be a significant time lag for every operation of the humanoid, making them unsuitable for fast and real-time applications such as touch lens manufacturing, electronic component inspection, sorting, packaging, and many other manufacturing industries as well as other non-manufacturing applications.

[0008] Although significant progress has been made in AI-based humanoid robots, they still face challenges. There is still a need to develop robots that can navigate complex environments, interact naturally with humans, and handle unpredictable situations. However, ongoing research and technological advancements continue to drive the development of more complex and capable AI-based humanoid robots.

[0009] The need for intensive data processing requirements remains a continuous challenge, so a straightforward solution is to provide a coordinated robot control system with high computational power in humanoid robots to ensure fast and real-time responses. Notably, technology continues to advance, and future generations of humanoid robots will require on-board computational power. With significant improvements in semiconductor miniaturization and processing power efficiency, it is now possible to incorporate multiple advanced digital signal processors (DSPs) that can handle dynamically allocated tasks, enabling the design of high-speed and highly intelligent humanoid robots. Summary of the Invention

[0010] In view of the background, an object of one aspect of the present invention is to provide a powerful coordinated robot control technology, including an on-board high-power AI computer module such as the NVIDIA AGX series module, capable of processing large amounts of data from multiple input sources to perform different tasks in parallel using a dedicated processor group within the AI computer module, rather than communicating with an external computer on the cloud or a server. In particular, this technology can enable the utilization of on-board processors to implement artificial intelligence and deep learning of humanoid robots to perform a set of tasks without relying on external resources where it is difficult to achieve real-time responses.

[0011] Another object of the present invention is to provide an industrial humanoid robot with on-board intelligence, which can process information, make autonomous decisions, infer operation tasks based on commands and external interface inputs, decrypt the in-operation images captured by a camera, and quickly execute tasks without sending information to an offline server and waiting for a response. The intelligence of the robots is enhanced by implementing reinforcement learning, in which they receive feedback in the form of rewards or punishments based on their responses to each command. Over time, the humanoid robots improve their decision-making processes to achieve better results in the form of accuracy, speed, consistency, and reliability among many other operational functions.

[0012] Another object of the present invention is to provide an ADCH system for integration into an in-line automation system associated with a specific process in a manufacturing environment, where a sophisticated robot is not required.

[0013] Another object of the present invention is to enable humanoid robots to be more autonomous and adaptive, thus allowing them to operate in various environments and efficiently execute complex tasks through deep learning techniques. Advanced robots utilize multiple types of sensors, such as cameras, remote sensing units, radars, transducers, and so on, to collect data about their environment. Sensor fusion techniques combine data from such sensors and other external interfaces, enabling the robots to have a more comprehensive understanding of their surrounding environment, which enables them to improve their AI responses over time.

[0014] Another object of the present invention is to enable humanoid robots to be more autonomous, where the motion control of the robots will be quickly computed to move the end effector of the robots to a specified destination. This is achieved by an on-board computing processor with inverse kinematics implemented to ensure smooth, accurate, and jerk-free movement of the robot arm and to enable anthropomorphic features.

[0015] Another object of the present invention is to combine a humanoid robot with more than one dedicated on-board computing processor (e.g., Jetson AGX Xavier series), so that all data analysis, data processing, and other real-time algorithms can be distributed on a parallel system architecture incorporated into the robot to ensure ultra-high-speed responses to specific commands from module-specific attachments and device-specific attachments. The goal is to ensure that these robots can respond to their environment and execute tasks in real time, thus adapting to changing conditions and interactions.

[0016] Another object of the present invention is to achieve rapid communication of a large amount of data (e.g., high-resolution images) between different nodes or processors via a USB C and an Ethernet star network, without being limited by the speed of a common bus through which they can communicate with each other via their boards. Such a rapid data communication strategy enables software applications to create maps of a manufacturing environment using sensor inputs. These maps can be used to identify obstacles and plan collision-free and shortest trajectory paths by combining techniques such as SLAM (Simultaneous Localization and Mapping) to determine the position and orientation of a robot in the mapped environment. Many analyzed trajectory paths can be further stored and used as part of predefined behaviors or movements to perform specific tasks, enabling autonomous navigation of a robot in indoor and familiar environments. The use of multiple processors interconnected by multiple data communication networks such as USB C, Ethernet, and a common bus architecture allows for extremely rapid data analysis based on real-time sensor data and continuous updating of planned trajectory paths to adapt to changing environmental conditions and avoid newly occurring dynamic obstacles.

[0017] Another object of the present invention is to facilitate efficient data transfer for distributed real-time imaging and inference across multiple nodes, thereby eliminating the common bottlenecks of latency-causing responses and inefficient performance encountered in conventional multi-GPU server systems that rely on a shared bus architecture.

[0018] Advances in robotics and artificial intelligence (AI) have made it possible to create humanoid robots with increasingly complex capabilities to perform complex tasks that require dexterity and precision. These robots typically combine technologies such as computer vision, natural language processing, machine learning, and sensor systems to perceive the world around them and interact with it, resulting in intelligent anthropomorphic robots.

[0019] With the growing demand for intelligent robots capable of performing complex tasks, there is a need for on-board intensive computing requirements to enable such robots in applications such as self-driving vehicles that can navigate on roads and avoid obstacles without human intervention, industrial robots deployed in cleanroom environments, robots assisting in surgeries in healthcare, rehabilitation, and other hospital duties where an extremely hygienic environment may be required to prevent the risk of infection transmission. They can also be deployed as agricultural robots for planting, harvesting, monitoring, and packaging, search and rescue operations to locate disaster areas such as floods, earthquakes, landslides, and other disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will be better understood by reading the following description of non-limiting embodiments and referring to the accompanying drawings: 6]

[0021] Figure 1 is a block diagram showing an example of an advanced distributed computing hardware system for coordinated robot control that can be used in a humanoid robot.

[0022] Figure 2 Shows the configuration of a single-board advanced distributed computing hardware, including four processors, each processor including a GPU (Graphics Processing Unit), an ARM-based CPU (Central Processing Unit), a VPA (Visual Processing Accelerator), and a DLA (Deep Learning Accelerator), and various types of interfaces for handling different tasks and controlling and monitoring the functions of a humanoid robot.

[0023] Figure 3 Is a block diagram showing a three-board advanced distributed computing hardware system, which shows multiple GPUs that control multiple interfaces through a network of communication modules and a common bus installed on the main board.

[0024] Figure 4 Shows a plan view of the single-board advanced distributed computing hardware configuration.

[0025] Figure 5 Shows a side view of the three-board advanced distributed computing hardware configuration.

[0026] Figure 6 Shows Figure 5 an isometric view of the three-board advanced distributed computing hardware system of

[0027] Figure 7 Shows a front view of a humanoid robot with built-in advanced distributed computing hardware.

[0028] Figure 8 Shows an isometric view of a humanoid robot with built-in advanced distributed computing hardware.

[0029] The drawings are not necessarily to scale and may be illustrated by graphical representation and sectional views. In some cases, details that are not necessary for understanding the embodiments or that make other details difficult to understand may have been omitted. Detailed Description

[0030] Describes an advanced distributed computing hardware (ADCH) system that is implemented using peer-to-peer communication between a cluster of processors that make up an action server, each processor including a GPU (Graphics Processing Unit), an ARM-based CPU (Central Processing Unit), a VPA (Visual Processing Accelerator), and a DLA (Deep Learning Accelerator) to perform various tasks. In particular, the ADCH system can allow coordinated processing to be distributed by Figure 3 the main processor 100 in Figure 3The tasks of 101 - 111) therein are to achieve the real - time functions of the robot. The network of processors can complete various coordinated human - like complex and multi - faceted tasks, or even simple coordinated tasks in real time through parallel processing and preemptive services. Such preemptive tasks and many other established processes are stored in the non - volatile memory within the robot and are further refined and enhanced by using machine learning and deep - learning algorithms.

[0031] One embodiment of the present invention uses a set of boards, including a cluster of 12 processors (such as those from the NVIDIA Jetson AGX series) 100 to 111, as Figure 1 shown. One Ethernet port from each processor is routed to an Ethernet switch 26 to form a star network topology. A peer - to - peer network is established between the processes running on each pair of processors ( Figure 1 100 to 111) in the cluster, and they communicate using different modes (such as broadcast messages, services, and action servers).

[0032] Broadcast messages can be published by any process in the network without knowing the subscribers of the message. It is a typical many - to - many connection for continuous data streams.

[0033] Services are implemented through short - duration remote service calls executed sequentially, otherwise known as synchronous service calls. During the execution of a remote service call, the scheduled call will remain active and will not be preempted by another remote service call. Thus, a typical service thread remains dedicated to a single service call until completion.

[0034] Action servers and clients establish a tight coupling of two or more processes so that different tasks are run by the assigned server, with the option of providing feedback during and at the end of the execution of service calls or requests from the client. It is important to note that action servers are designed to be preemptive and non - blocking, which means they are capable of multitasking.

[0035] All processes are designed to be fine - grained and modular, such that each process performs a well - defined task and has a callable interface. These interfaces are exposed to any of the above - mentioned communication modes (broadcast messages and remote service calls) based on the required processes.

[0036] A reliable safety system is integrated into the ADCH. These systems include collision detection by leveraging infrared and proximity sensors, an emergency stop mechanism, and a fail-safe to prevent damage to the robot or its surrounding environment, as well as self-diagnostic features that detect and report faults within the computer and errors or inappropriate responses from external hardware interfaces in real time. Additionally, password-protected data security and privacy are ensured within the ADCH, which enables data to be processed and stored locally without the risk of data exposure during transmission to external servers, especially in cases involving sensitive applications.

[0037] A modular process is deployed in all twelve processors ( Figure 1 from 100 - 111), and a main processor 100 is capable of indicating which process is assigned to which processor. To train, configure, program, and call any tasks that may be required to perform an action, the main device 100 is connected to the Figure 1 monitor and mouse 10 via an HDMI port. Given the open nature of the processes and network topology, the main device has the ability to dynamically assign specific processors to specific processes based on the required processing power and the processor load at any stage. In effect, the main device manages the cluster like an operating system (OS) scheduler that manages multi-core processors.

[0038] Reference Figure 1 , the ADCH system 200 includes three boards B1, B2, and B3, each board consisting of four Xavier processors. A cluster of a total of twelve processors ( Figure 1 from 100 to 111) is available to perform multiple tasks in parallel through high-speed communication via a dedicated PCI bus on the main board 30. Due to space constraints, Figure 1 not all Xaviers are shown in Figure 1 In the embodiment shown, 100 is the main device that manages the process allocation of processors 101 - 111. External interfaces (such as motors 20, cameras 16 and 14, sensors (not shown) via motor driver 18) are connected to Figure 1Another processor 103 therein. All Ethernet and I2C connections from each processor terminate at the BUS to connect to the motherboard 30 via the BUS connector 22. Inter-board communication can be implemented to work between peers or via a high-speed bus interconnecting three processor boards B1, B2, and B3. The motherboard 30 also provides general input / output connections 23 via the I2C I / O expander 24 and communication to the Internet 28 via the Ethernet switching module 26. Power is provided by 21. Since cameras 14 and 16 require high data speeds, they dock with a USB C port that provides a peak speed of up to 40 Gbps, especially when using high-resolution cameras, which enables real-time data transfer. The USB A port is used for data transfer speeds below 10 Gbps for cases where it is sufficient for the attached accessories. The 3BUT is an I / O port where sensors and / or switches (not shown) are connected to enable action status feedback, hard reset, and any other hardware input / output configuration options that may be required.

[0039] The ADCH system 200 operating system is programmed to configure the twelve processors to perform different types of tasks. For example, in Figure 1 the illustrated embodiment, processors 100 and 101 are assigned as control nodes, 102 and 103 as planning and sequencing nodes respectively, 104 and 105 as user interface and debugging nodes, and the remaining processors (processors 106 and 111) are assigned as imaging nodes. The ADCH system software application provides the flexibility to dynamically change node assignments based on processor load to accelerate data processing, thus achieving a high degree of real-time performance. The system is also designed to scale up performance by adding more processors or reassigning processors for different tasks.

[0040] The ADCH system is built into the robot, and all operations are distributed through nodes that operate in a multi-tasking environment providing a wide parallel processing environment. The nodes are also referred to as processors, each of which is connected to a non-volatile memory that enables the storage of the state conditions of a specific sequence to ensure easy recovery of robot movement in case of a shutdown or power-off condition through a well-coordinated control mechanism using a star network. Traditional control systems require a long recovery process starting from communicating with an external server, waiting for the state of the last known condition of the machine or robot to restart the robot from a specific original position, followed by a command to restore the machine or robot to the last known position. The nodes can also store the results of data analysis of all or any configurations (Viz... image inspection), which can be used by algorithms to build an artificial intelligence database and enhance machine learning. Nodes 100 and 101 are used as control nodes. They allocate tasks, track the status of each other's processors and their respective computational loads at any given time, and deploy neural networks through any other idle nodes 102 to 111. Due to the nature of the real-time requirements of the robot, a feature is implemented where software applications can control whether the processing is performed by software or hardware, thus ensuring a quick response. The nodes also manage the power consumption of the system by turning off the clocks to unused nodes, thus effectively reducing the heat dissipation of the robot. Nodes 102 and 103 respectively manage the planning and sequencing of the ADCH external interface with full-duplex functionality and ensure the optimal use of computational capabilities. The planning and sequencing control nodes assist in operating the robot's movement in the smoothest and fastest trajectory path by the optimal calculation of robot joint angles and the optimal trajectory planning to maintain balance and avoid collisions with intermediate obstacles. Nodes 104 and 105 control the user interface (UI) and debugging operations respectively during the training and configuration of the robot. Nodes 106 to 111 are dedicated to visual analysis, including imaging, trajectory planning, balance control through optimal robot joint angles, algorithms for inverse kinematics, processing, and feedback of the results to the node network and the main node 100. The main node 100 in ADCH can allocate any node for audio support when needed, where the robot can recognize audio commands (speech recognition), speech recognition, natural language processing, and decision-making, enabling the ability to understand natural language commands and generate appropriate natural language responses. Machine learning and deep learning techniques are often used to improve the capabilities of robots in these areas and make corresponding responses to execute tasks or task sets. In addition, feedback in the form of speech responses makes the robot scalable and flexible to adapt to various applications. Advanced audio features can be implemented in cases where the robot needs to understand and analyze multi-language commands to implement them in various countries without the user having to operate the robot using a specific language UI (user interface).

[0041] Figure 2Shows a typical board layout of a single computing board B1, also known as a "System on Module (SOM)", which includes four external processors or nodes. The external interfaces are terminated on one side of the board and communicate the interfaces to the main board 30 through a connector 22. Some of the interfaces relevant to the present invention are on-board Wi-Fi, Bluetooth, general-purpose I / O, Ethernet, or other communication interfaces to interact with people or other devices. This connectivity allows them to receive commands and send information. The general-purpose I / O (GPIO) is provided through an I2C interface board 24 and an Ethernet communication interface via an Ethernet switch 26. Interface 12 is a host flash memory system that plays an important role in configuring the ADCH system by enabling functionality such as uploading / downloading firmware, configuration, and uploading / downloading of all other relevant operating parameters. The host flash computer 12 plays a key role in establishing an ADCH system integrated with a software development tool library module to add / modify the operating system kernel of the processor(s) and customize software applications, boot loaders, and device drivers to perform a specific task set for processing different products, making the ADCH scalable and flexible to adapt to changes in robotic applications. The host flash computer 12 can also be used to flash other ADCH systems that control other robots performing the same set of operations through a process called mirroring or cloning, so that the optimized programs of the robots can be replicated and cloned to operate another robot to ensure a stable and consistent operating environment.

[0042] Figure 3 Shows a block diagram of a typical ADCH system that includes three computing boards B1, B2, and B3 installed on the main board 30 through a slot connector 22. Figure 3 Also shows how common signals are docked between multiple processors (such as 100 - 111) to achieve a powerful and efficient system for monitoring and controlling various devices, such as Optispec cameras 16 and 19, printers 13, BO module 15, robotic head camera 14, and a motor controller called Elmo 18. The main board provides access to the Internet through an Ethernet switching module 26 and access to input / output signals through an I2C GPIO expander 24 to communicate with external devices. Figure 3 The configuration shown is designed to be of a small size (form factor) to be able to fit into a robot that has sufficient built-in data processing capabilities and is very small in terms of the requirement to communicate with an external server to meet its computing needs. A scalable robotic environment is the possibility with such a distributed and self - contained system.

[0043] Figure 4 、 5 Figures 4, 5, and 6 show a plan view, a side view, and an isometric view of the ADCH system that can be easily deployed inside a robot.

[0044] Figure 7 and 8 shows a front view and an isometric view of a robot incorporating an ADCH in this embodiment of the present invention. Various devices installed within the robot are not shown as they fall outside the scope of the present invention.

[0045] Reference has been made to the accompanying drawings and specific language has been used herein to describe the same. However, it should be understood that no limitation of the scope of the present technology is thereby intended. Variations and further modifications of the control system features shown herein should be considered to be within the scope of this specification.

[0046] However, it will be recognized that the present technology may be practiced without one or more of the specific details, or with other methods, hardware components, interface devices, etc. Well-known modules such as upload and download processing or operations are not shown or described in detail so as not to obscure aspects of the present invention.

[0047] The subject matter defined in the appended claims need not be limited to the specific features and operations described above. Rather, the above specific features are disclosed as example forms for implementing the claims. Many modifications or arrangements may be devised without departing from the spirit and scope of the described invention.

Claims

1. A coordinated control system designed to have a small form factor, the coordinated control system being capable of being located within a robot or an automated on-line manufacturing system, the coordinated control system comprising: On-board advanced distributed control hardware configured with an artificial intelligence-enabled processor incorporated within a "System on Module (SOM)" board, each SOM including four processors or nodes; Docking with cameras, motors, general-purpose I / O via an I2C expander, an audio interface, the Internet, wireless communications such as Wifi and Bluetooth for sending and receiving commands and any other data; Electrically connected to a plurality of sensors for sensing the environment and collecting data from infrared, tactile, proximity, and other types of sensors; An external host flash computer dedicated to uploading / downloading firmware, cloning, and configuration settings; Non-volatile memory for storing control algorithms, configuration files, and other necessary data; An HDMI-based user interface for robot operation, training, and setup; An internal star network based on USB C or Ethernet for high-speed communication bypassing the standard PCI bus interface BUS; An Ethernet switch enabling multiple boards to access the Ethernet for internal communication within the star network and external Internet access.

2. The coordinated control system according to claim 1, further comprising: An artificial intelligence-enabled system for local data processing, enabling a significant reduction in latency in making autonomous decisions supported by artificial intelligence-based machine learning and deep learning algorithms; A user interface for robot interaction, wherein the HDMI-based user interface includes a display, a mouse for visual and voice communication with a human operator; A plurality of processors, each processor comprising a GPU (Graphics Processing Unit), an ARM-based CPU (Central Processing Unit), a VPA (Visual Processing Accelerator), and a DLA (Deep Learning Accelerator) for analyzing the environment and achieving real-time performance.

3. The coordinated control system according to claim 1, further comprising: A power management function within the ADCH for optimizing power or battery usage by effectively managing the processors and disabling unused processor clocks to ensure low heat dissipation.

4. The coordinated control system according to claim 1, further comprising: A self-diagnostic system for real-time detection and reporting of faults within the computer, as well as errors or inappropriate responses from external hardware interfaces.

5. The coordinated control system according to claim 1, further comprising: Data security and privacy functions enabling local processing and storage of data without the risk of data exposure during transmission, which is crucial for sensitive applications.

6. The coordinated control system according to claim 1, further comprising: A host flash computer dedicated to modifying or enhancing the operating system kernel of the plurality of processors to customize the software applications, bootloader, and device drivers that contribute to scalability and flexibility.

7. The coordinated control system according to claim 1 also enables the implementation of: A distributed and self-contained robot environment that is scalable and suitable for new applications.

8. A coordinated control method for a robot or an automated online system in a manufacturing environment, the method comprising: Utilizing the built-in computer in the ADCH to coordinate and execute tasks related to the production process, the tasks including material handling and quality control; Receiving external production instructions via the WiFi, Bluetooth or Ethernet from a central manufacturing control system or from a software application residing in the built-in computer; Fast communication within the ADCH implemented by broadcast messages, the broadcast messages being able to be published by any process in the star network without any knowledge of the subscribers of the messages, resulting in a typical many-to-many connection for continuous data streams; A peer-to-peer network established between processes running on each pair of processors; Management of synchronous service calls achieved by sequentially executed short-duration remote service calls, remaining dedicated and active during the execution of the short-duration remote service calls and not being preemptible by another remote service call; Options for scaling the functionality and speed of the ADCH by reallocating unused nodes or processors, contributing to scalability and flexibility; Dynamic allocation of nodes or processors by the master node for efficient task allocation during normal operation, debugging and user interface utilization during training and configuration, and during data and image analysis to obtain maximum computing speed; Action servers and clients for establishing a tight coupling of two or more processes to run different tasks by the allocated servers, with the option of providing feedback during or at the end of the execution of a remote service call or service request from the client, and wherein the action servers are designed to be preemptible and non-blocking, such that they can perform multiple tasks with password-protected data security and privacy functions within the ADCH to locally process and store data, without the risk of data exposure during transmission to an external server when used in sensitive applications.

9. The coordinated control method according to claim 8, wherein: Processes are designed to be fine-grained and modular to ensure that each process performs well-defined tasks with callable interfaces.

10. The coordinated control method according to claim 9, wherein the callable interfaces are exposed by the processes to all communication modes (broadcast messages and remote service calls) based on the requirements.

11. The coordinated control method according to claim 8, wherein the sequence of robot movement is analyzed and calculated for the robot joint angles for balance management.

12. The coordinated control method according to claim 8, wherein the trajectory path of the robot is planned by implementing an algorithm for inverse kinematics to ensure smooth, jerk-free and accurate movement capable of achieving anthropomorphic characteristics.

13. The coordinated control method according to claim 8, wherein predefined behaviors or movements are utilized to perform specific tasks for autonomous navigation of a robot in indoor and familiar environments.

14. The coordinated control method according to claim 8, wherein an effective AI algorithm is implemented for object recognition, speech recognition, natural language processing, and decision-making, such that natural language speech commands are understood and appropriate natural language responses are generated.

15. The coordinated control method according to claims 8 and 14, wherein the vision system contributes to the understanding of the manufacturing environment and quality inspection capabilities, which are supplemented by artificial intelligence inference, deep learning, and reinforcement learning for efficient end-to-end autonomous applications.

16. The coordinated control method according to claim 8, wherein efficient data transfer is facilitated through the Ethernet and USB star network, as well as the interfaces within the ADCH, to enable distributed real-time image processing and inference across multiple nodes, and to overcome common bottlenecks encountered in conventional multi-GPU server systems that rely on a shared bus architecture.

Images