Overall design scheme of water-land-air triphibian multifunctional intelligent robot based on artificial intelligence technology

Abstract

The invention discloses an overall design scheme of a water-land-air triphibian multifunctional intelligent robot based on an artificial intelligence technology, and the robot has excellent environment adaptability, and a multi-mode environment sensing system and an intelligent decision planning system are closely cooperated, so that the robot can rapidly adapt to complex changes of different environments of water, land and air, and the robot is more intelligent. Environment information is accurately sensed, a reasonable decision is made, and the working capacity in a complex environment is remarkably improved; the robot also has strong intelligent decision-making ability, a reinforcement learning algorithm and a continuous learning mechanism endow the robot with strong autonomous decision-making ability, so that the robot can quickly formulate and adjust action strategies when facing complex tasks and emergencies, and the task execution efficiency and success rate are effectively improved; the robot is provided with efficient energy utilization, diversified energy supply and an intelligent energy management system, the energy utilization efficiency is effectively improved, the endurance time of the robot is prolonged, and it is ensured that the robot can continuously and stably work in long-time and long-distance tasks.

Description

Technical Field

[0001] This invention deeply integrates robotics and artificial intelligence, focusing on creating a multifunctional intelligent robot capable of amphibious operations on land, water, and in the air. Leveraging cutting-edge AI technology, this design aims to comprehensively enhance the robot's adaptability and mission performance in complex environments. It is widely applicable in a variety of key areas, including environmental monitoring, emergency rescue, scientific research, and logistics. Background Art

[0002] As global science and technology continue to advance at a rapid pace, humanity's demand for exploring and operating in complex environments is exploding. In the field of environmental monitoring, accurate and real-time monitoring of marine ecological changes, forest resource distribution, and high-altitude meteorological data is crucial for ecological protection and sustainable development. In emergency rescue scenarios, the frequent occurrence of natural disasters such as floods, fires, and earthquakes requires rescue equipment to quickly reach affected areas and effectively carry out rescue operations. In scientific research and exploration, the mysteries of unknown deep-sea creatures and the unique geological structures in remote mountainous areas urgently require advanced tools capable of operating across land, water, and air. However, traditional single-environment operating equipment is increasingly exposed to numerous limitations when faced with these complex and diverse tasks, making it difficult to meet practical needs. The rapid development of robotics and artificial intelligence technologies has created new opportunities to address these challenges, making the development of multifunctional intelligent robots capable of operating on land, water, and air a highly promising research direction.

[0003] Currently, robots for land, water, and air environments have achieved varying degrees of development. Land robots are widely used in industrial production and logistics. For example, handling robots on industrial production lines, through precise programming and control, can efficiently and accurately move a wide range of goods. Sorting robots in logistics warehouses, leveraging advanced visual recognition and path planning technologies, significantly improve cargo sorting efficiency. Aquatic robots are primarily used in scenarios such as ocean monitoring and water patrols. Unmanned monitoring boats equipped with specialized water quality monitoring equipment can provide real-time monitoring of water quality, flow, and other parameters over large areas of water, providing robust data support for water resource protection. Water patrol robots can effectively perform security patrols and ensure water safety. Aerial robots, represented by drones, play an important role in aerial photography, agricultural plant protection, and power inspections. Aerial drones can capture high-resolution images and videos, providing unique perspectives for urban planning, film and television production, and more. Agricultural drones improve agricultural production efficiency while reducing labor costs through precise pesticide spraying. Power inspection drones can quickly monitor the operating status of power lines and promptly identify potential faults.

[0004] Despite progress in existing robotics technology, a core and pressing issue remains: the lack of a universal intelligent robotic solution capable of efficiently switching between land, water, and air environments and operating collaboratively. Most robots currently available are limited to a single environment and lack the flexibility to adapt to changing circumstances, severely limiting their scope of application. For example, in complex post-earthquake rescue scenarios, land-based rescue robots struggle to reach the disaster site across water, aquatic rescue equipment struggles to maneuver on land, and aerial robots, while capable of reconnaissance, struggle to execute rescue operations. Furthermore, existing robots lack the intelligence required to make decisions autonomously in complex and changing environments, often requiring extensive human intervention. This not only increases labor costs but also reduces the efficiency and accuracy of task execution, making it difficult to meet the diverse demands of complex scenarios. Summary of the Invention

[0005] This invention is committed to breaking through the bottleneck of existing technologies and providing an innovative overall design scheme for an amphibious, multifunctional intelligent robot based on artificial intelligence technology. By integrating advanced artificial intelligence algorithms, multimodal perception technology, efficient energy management systems, and diversified communication and data processing technologies, the robot is given the ability to operate autonomously in land, water, and air environments. This enables it to quickly and accurately adapt to environmental changes, make scientific and reasonable decisions, and efficiently complete various complex tasks, thereby expanding the application boundaries of robots in multiple fields and providing strong technical support for the intelligent development of related industries.

[0006] The artificial intelligence-based humanoid robot dynamic environment adaptation and behavior optimization method of the present invention mainly includes the following core steps:

[0007] S1: Multimodal environmental perception system: The robot is equipped with a variety of advanced sensors, including high-definition cameras, infrared cameras, lidars, ultrasonic sensors, pressure sensors, and humidity sensors. A multimodal data fusion algorithm based on deep learning is used. During the data fusion process, the visual sensor data is V, the lidar data is L, the ultrasonic sensor data is U, the pressure sensor data is P, and the humidity sensor data is H. The fused data is F, and the fusion is performed using the following general formula:

[0008] F = w1f1(V)+w2f2(L)+w3f3(U)+w4f4(P)+w5f5(H), where w1w2w3w4w5 are the weights of each sensor data, and f1f2f3f4f5 are the functions for feature extraction and conversion of each sensor data. By adjusting the weights and functions, accurate perception of objects, terrain, weather, and other information in different environments can be achieved. For example, in the ocean environment, visual sensors identify marine life and obstacles, pressure sensors monitor sea depth, and lidar detects underwater terrain. By fusing these data through algorithms, a comprehensive and accurate ocean environment model is constructed.

[0009] S2: Intelligent decision-making and planning system: Use reinforcement learning algorithms, combined with environmental perception information and task objectives, to formulate the optimal action strategy for the robot; in reinforcement learning, taking the deep Q network (DQN) as an example, the calculation of Q value follows the following formula: Q(s t ,a t )=r t +γmax at+1 Q(s t+1 ,a t+1 ) where Q(s t ,a t ) means in state s t Take action t Q value, r t is to take action t The reward obtained later, γ is the discount factor, which is used to balance the importance of current rewards and future rewards, s t+1 is to take action t The next state after a t+1 It is the possible action to be taken in the next state; the robot learns the optimal behavior strategy by continuously iteratively calculating the Q value;

[0010] In terms of path planning, for simple environments, the optimized A algorithm is used; the heuristic function of the A algorithm usually adopts a variant of Manhattan distance or Euclidean distance, and the general form can be expressed as: h(n) = c d(n, goal)

[0011] Here, h(n) is the heuristic estimate of the distance from node n to the target node, d(n, goal) represents a distance metric (such as Manhattan distance or Euclidean distance) from node n to the target node, and c is the adjustment coefficient. By properly setting the heuristic function, the A* algorithm can quickly plan an initial path. For complex dynamic environments, the Dijkstra algorithm and the dynamic window algorithm (DWA) are integrated to dynamically adjust the path based on real-time environmental changes, ensuring that the robot can safely and efficiently avoid obstacles and reach the target location. For example, in a fire rescue scenario, the robot uses this system to plan the optimal rescue path based on information such as the direction of fire spread, building structure, and personnel distribution.

[0012] S3: Artificial Intelligence Algorithm Optimization and Learning System: Through continuous online learning and transfer learning mechanisms, the robot continuously optimizes its decision-making model. During online learning, the robot dynamically adjusts algorithm parameters based on real-time environmental data and task execution feedback to adapt to dynamic environmental changes. Transfer learning helps the robot quickly adapt to new environments, transferring knowledge learned in one environment to other environments. For example, when the robot switches from freshwater environmental monitoring to seawater environmental monitoring, the transfer learning mechanism allows it to quickly adapt to the characteristics of the seawater environment with only a small amount of data fine-tuning.

[0013] S4: Energy Management and Power Systems: Design an efficient energy management system, using a variety of energy supply methods such as rechargeable batteries, solar panels, and hydrogen energy. Achieve efficient energy utilization through intelligent energy switching strategies based on environmental and mission requirements. Design propulsion methods tailored to different environments, using propellers or water jets in water, electric propellers or jet propulsion in the air, and wheeled, tracked, or foot-powered propulsion on land, depending on the terrain. For example, during long-term ocean monitoring missions, solar power can be used for charging during the day and batteries can be used for power at night. When operating on land in complex terrain, the propulsion method can be automatically switched based on terrain changes.

[0014] S5: Communication and data processing system: Build a stable and reliable communication network that supports satellite communication, 5G communication, and ad hoc network communication; adopt a combination of cloud computing and edge computing to quickly process data with high real-time requirements locally through edge computing devices to reduce transmission delays; upload large amounts of data to the cloud computing platform for in-depth analysis and storage to achieve efficient data management and utilization; for example, in environmental monitoring tasks, edge computing devices analyze sensor data in real time and detect anomalies; the cloud computing platform performs trend analysis and modeling on long-term monitoring data to provide strong support for decision-making.

[0015] In summary, the beneficial effects of the present invention are:

[0016] 1. The present invention has excellent environmental adaptability. The multimodal environmental perception system and the intelligent decision-making and planning system work closely together, enabling the robot to quickly adapt to the complex changes in different environments on land, water and air, accurately perceive environmental information and make reasonable decisions, significantly improving its ability to operate in complex environments.

[0017] 2. The present invention has powerful intelligent decision-making capabilities. The reinforcement learning algorithm and continuous learning mechanism give the robot powerful autonomous decision-making capabilities, enabling it to quickly formulate and adjust action strategies when faced with complex tasks and emergencies, effectively improving the efficiency and success rate of task execution.

[0018] 3. The present invention has efficient energy utilization, diversified energy supply and intelligent energy management system, which effectively improves energy utilization efficiency, extends the robot's endurance time, and ensures that the robot can continue to operate stably during long-term and long-distance tasks.

[0019] 4. The present invention has broad application potential and can be applied to multiple fields such as environmental monitoring, emergency rescue, scientific research, logistics and distribution, etc. It can provide innovative solutions for various industries and promote the intelligent development process in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 This is a schematic diagram of the core architecture of the robot of the present invention;

[0022] Figure 2 This is a schematic diagram of the process from multimodal data acquisition to decision-making in the present invention;

[0023] Figure 3 This is a schematic diagram of the intelligent decision-making logic flow of the present invention.

[0024] Figure 4 Schematic diagram of the closed-loop management process of the energy and power system of the present invention. DETAILED DESCRIPTION

[0025] The following is a combination of the embodiments of the present invention Figure 1-4The technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0026] Example

[0027] An overall design scheme for an amphibious multifunctional intelligent robot based on artificial intelligence technology includes the following implementation steps:

[0028] Implementation of multimodal environmental perception system

[0029] Sensor selection and layout: Sensors are carefully arranged on the robot's surface; high-definition cameras and infrared cameras are installed on a 360-degree rotating pan-tilt device to ensure a wide field of view; a lidar is installed on the top of the robot to achieve 360-degree scanning of the surrounding environment; ultrasonic sensors are evenly distributed around the robot for close-range obstacle detection; pressure sensors are installed in key locations inside the robot to monitor water or air pressure according to different environments; and humidity sensors are installed in a well-ventilated location that can accurately sense changes in ambient humidity.

[0030] Data fusion algorithm implementation: A fusion algorithm based on CNN and LSTM is adopted. First, CNN performs multi-layer convolution and pooling operations on visual sensor data to extract key features in the image. LSTM processes time series data such as pressure and humidity to capture the temporal dependency of the data. Then, feature data from different sensors are fused through a specific fusion strategy, and finally classified and regressed through the fully connected layer to achieve accurate perception of environmental information. For example, in forest environmental monitoring, the fusion of visual and humidity sensor data can accurately determine the risk level of forest fires.

[0031] Implementation of intelligent decision-making and planning systems

[0032] Reinforcement learning algorithm training: Clearly define the robot's state space, action space, and reward function; the state space covers environmental perception information and the robot's own state (position, speed, posture, etc.); the action space includes the robot's possible actions in different environments (such as forward, backward, turning, ascending, descending, etc.); the reward function is designed according to the mission objectives and environmental constraints. For example, in a rescue mission, successful rescue of the target is rewarded with a high reward, while approaching a dangerous area is punished; use the deep Q network (DQN) or the proximal policy optimization algorithm (PPO) for training to enable the robot to learn the optimal behavioral strategy.

[0033] Application of path planning algorithms: When planning a path, the environmental perception data is first used to build a map model. For simple environments, the optimized A* algorithm is used to quickly search for the initial path through a heuristic function. In complex dynamic environments, the Dijkstra algorithm and the dynamic window method (DWA) are combined to monitor environmental changes in real time, such as the movement of obstacles and the emergence of new obstacles, and the path is dynamically adjusted to ensure that the robot can reach the target location safely and efficiently. For example, in urban fire rescue, the rescue path is dynamically adjusted by comprehensively considering the risk of building collapse, the direction of fire spread, and road congestion.

[0034] Artificial intelligence algorithm optimization and learning system implementation

[0035] Online learning mechanism: During operation, the robot records every decision and environmental feedback to form empirical data. At regular intervals, this new data is used to fine-tune the reinforcement learning model and update the parameters of the neural network. For example, when the robot operates in a new water environment, it continuously optimizes its response strategy to water flow speed, direction, and underwater obstacles through online learning.

[0036] Transfer learning applications: Before a robot enters a new environment, it is pre-trained in a simulated environment, using the model parameters learned in a similar environment as initial values. Then, it is fine-tuned in the actual environment using a small amount of data to quickly adapt to the new environment. For example, when switching from freshwater environment monitoring tasks to seawater environment monitoring, transfer learning can be used to accelerate the adaptation process to the seawater environment and reduce the time and cost of retraining.

[0037] Energy management and power system implementation

[0038] Energy supply module configuration: High-efficiency solar panels are installed on the surface of the robot, and lithium-ion batteries or hydrogen fuel cells with high energy density and long cycle life are selected as the main energy storage devices; for water operations, a waterproof and efficient hydrogen power generation device is equipped; when operating on land, motor drive systems of different power are selected according to the terrain and task requirements, such as wheel drive on flat terrain and track drive on rugged terrain; lightweight, high-power electric propellers or jet engines are used for aerial operations.

[0039] Energy management strategy: Develop an intelligent energy management strategy based on remaining energy, environmental energy availability, and mission requirements. Prioritize solar charging when sufficient solar energy is available. When the mission is urgent and energy is insufficient, switch to a high-energy-density energy supply mode (such as hydrogen fuel cells). For example, during long-term ocean monitoring missions, fully utilize solar energy for charging during the day and use batteries for power at night to ensure a stable energy supply.

[0040] Communication and data processing system implementation

[0041] Communication network construction: Integrate satellite communication modules, 5G communication modules and ad hoc network communication modules on the robot; use satellite communication to maintain contact with the control center in remote areas or areas without network coverage; use 5G communication to achieve high-speed data transmission in cities or areas with 5G network coverage; in multi-robot collaborative operation scenarios, achieve information sharing and collaborative control between robots through ad hoc network communication.

[0042] Data processing architecture: Edge computing devices, such as NVIDIA's Jetson series modules, are deployed locally on the robot to preprocess and perform preliminary analysis on the real-time data collected by the sensors and extract key features. Large amounts of historical data and complex data analysis tasks are uploaded to a cloud computing platform (such as Alibaba Cloud or Tencent Cloud), leveraging the powerful computing power of cloud computing for in-depth mining and storage management. For example, in environmental monitoring tasks, edge computing devices analyze sensor data in real time and detect anomalies, while the cloud computing platform performs trend analysis and modeling on long-term monitoring data to provide a basis for predicting environmental changes.

[0043] Comparison of the effects of the embodiments

[0044]

[0045] Through comparison of the above embodiments, it can be clearly seen that the robot of the present invention has significant advantages in environmental adaptability, intelligent decision-making ability, endurance and task completion rate, etc., and can effectively solve the problems existing in the existing technology and meet the diverse task requirements in complex environments.

[0046] It should be noted that the functions to be implemented by the hardware in the present invention are supported by a large number of mature technologies and belong to the existing technology. The essence of the present invention is to optimize the combination of existing hardware and its connection methods for specific application scenarios to meet the adaptation requirements in specific application scenarios and solve the problems raised in the background technology (not involving improvements to the software inside the hardware).

[0047] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology, characterized by: The following steps are involved: S1: Multimodal environmental perception system, used to collect and integrate various information from different environments such as land, water and air; S2: Intelligent decision-making and planning system, which formulates action strategies and path planning based on artificial intelligence algorithms; S3: Artificial intelligence algorithm optimization and learning system, enabling online learning and transfer learning of robots; S4: Energy management and power systems, using multiple energy supplies and adapting to different environments; S5: Communication and data processing system, which supports multiple communication methods and combines cloud computing and edge computing to process data.

2. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The multimodal environmental perception system includes multiple sensors such as vision, lidar, ultrasound, pressure, humidity, etc., and uses a fusion algorithm based on convolutional neural network and long short-term memory network for data fusion. The fusion formula is F=w1f1(V)+w2f2(L)+w3f3(U)+w4f4(P)+w5f5(H).

3. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The intelligent decision-making and planning system uses reinforcement learning algorithm to formulate action strategies, where the formula for calculating Q value of deep Q network is Q(s t ,a t )=r t +γmax at+1 Q(s t+1 ,a t+1 ), the improved A algorithm or Dijkstra algorithm combined with the dynamic window method is used for path planning. The general form of the heuristic function of the A algorithm is h(n) = c·d(n,goal).

4. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The artificial intelligence algorithm optimization and learning system includes an online learning mechanism and a transfer learning mechanism, which are used to optimize the robot's decision-making model.

5. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The energy management and power system adopts a variety of energy supply methods such as rechargeable batteries, solar panels, hydrogen energy, etc., and designs propulsion methods suitable for different environments.

6. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The communication and data processing system supports satellite communication, 5G communication and ad hoc network communication, and combines cloud computing and edge computing for data processing.

7. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The sensors in the multimodal environmental perception system adopt a specific layout to achieve all-round environmental information collection.

8. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The reward function in the intelligent decision-making and planning system is designed according to the task objectives and environmental constraints.

9. The overall design scheme of an amphibious multifunctional intelligent robot based on artificial intelligence technology according to claim 1 is characterized in that: The energy management and power system intelligently switches energy modes according to remaining energy, environmental energy acquisition conditions, and mission requirements.

Images