Intelligent nursing bed voice control system

Abstract

The application discloses an intelligent nursing bed voice control system and belongs to the technical field of intelligent nursing equipment. In view of the defects that the traditional nursing bed voice recognition only supports standard mandarin, the recognition rate of non-standard spoken language instructions is low, there is no secondary confirmation mechanism, and secondary injury of patients is easily caused, the application adopts a modular architecture, and contains a voice collection module, a big data voice recognition module, an AI intelligent verification module, an interaction feedback module, a permission control module, a safety limit determination module and a bed body driving control module. The application relies on a cloud big data voice library, adapts to non-standard mandarin and various spoken language instructions, and intelligently intercepts illegal and over-limit dangerous operations through AI semantic verification, man-machine bidirectional confirmation, hierarchical permission control and safety threshold double verification. The application is suitable for patients with various pronunciation habits, effectively reduces the risk of voice misoperation, avoids secondary injury such as bed falling and limb compression, has high recognition fault tolerance and high safety, and is suitable for hospital, old-age care institution and family intelligent nursing scenes.

Description

Technical Field

[0001] This invention relates to the field of intelligent nursing bed control technology, specifically to an intelligent nursing bed voice control system. Background Technology

[0002] As a core piece of equipment in the fields of health and wellness and medical care, smart nursing beds are widely used in hospital wards, elderly care institutions and home care scenarios. With the advantages of not requiring manual operation and being suitable for patients with limited mobility, voice control is gradually replacing traditional button and touch control methods and has become a standard feature of smart nursing beds.

[0003] Currently available intelligent nursing bed voice control systems have significant technical flaws: First, their voice recognition accuracy relies on standard Mandarin, only recognizing fixed phrases and standard pronunciation commands. This makes them unsuitable for patients with heavy accents, non-standard pronunciation, or casual, colloquial expressions. Elderly patients and those from dialect regions are particularly prone to command recognition failures and errors, resulting in extremely poor interactive compatibility. Second, most existing systems operate on a one-way command execution model, directly driving the nursing bed's movements after receiving voice commands. There is no secondary confirmation or real-time feedback mechanism, meaning invalid commands from patients' mispronunciations or casual expressions will be executed directly. Third, they lack intelligent access control and security mechanisms, failing to distinguish patient intentions or determine the safety boundaries of actions. When patients are in poor physical condition, confused, or issue commands exceeding limits, the nursing bed may experience dangerous actions such as significant lifting, tilting, or side-tipping, leading to secondary injuries such as falls, limb compression, or postural discomfort, posing a significant medical safety hazard.

[0004] In summary, existing intelligent nursing bed voice control technology suffers from low recognition error tolerance, insufficient intelligent interaction, and lack of safety protection. It cannot meet the convenient use needs of patients with different pronunciation habits and cannot guarantee the safety of nursing operations. There is an urgent need for a highly adaptable, highly secure voice control system with intelligent verification feedback. Summary of the Invention

[0005] This invention aims to address the technical pain points of existing intelligent nursing bed voice control systems, such as high reliance on standard pronunciation, weak recognition of colloquial commands, lack of command confirmation feedback, and lack of intelligent safety protection. It provides an intelligent nursing bed voice control system that accurately recognizes non-standard Mandarin and colloquial commands, confirms operation commands through real-time human-machine feedback, and combines AI intelligent permissions and safety limit determination mechanisms to balance ease of operation with patient safety, preventing secondary injuries caused by misoperation.

[0006] To achieve the above objectives, the present invention provides an intelligent nursing bed voice control system, including a voice acquisition module, a big data voice recognition module, an AI intelligent verification module, an interactive feedback module, an access control module, a safety limit determination module, and a bed drive control module; The voice acquisition module is located on both sides of the head of the smart nursing bed. It adopts a dual-microphone noise reduction array structure to collect patients' spoken voice commands in real time, filter environmental noise, breathing noise, and equipment operating noise, and output pure original voice data. The big data speech recognition module is electrically connected to the speech acquisition module and the AI ​​intelligent verification module respectively. The big data speech recognition module has a built-in cloud-linked multi-dimensional big data speech library. The big data speech library includes non-standard Mandarin pronunciations from multiple regions, colloquial and casual speech, dialect Mandarin variants, and fuzzy command expression data. It is used to extract features, correct semantics, and perform fuzzy matching on the collected non-standard and colloquial speech commands, and to convert the colloquial and non-standard original speech into standardized equipment control commands. The AI ​​intelligent verification module is electrically connected to the interactive feedback module. It is used to receive standardized control commands converted by the big data speech recognition module, and to perform intelligent verification by combining contextual semantics and the patient's historical operating habits. It filters out ambiguous commands, unclear commands, and invalid commands, and generates corresponding confirmation feedback messages. The interactive feedback module has a built-in voice broadcasting unit and a light prompting unit, which are used to receive feedback signals from the AI ​​intelligent verification module, repeat the instruction to be executed to the patient through voice broadcasting and request confirmation, and at the same time use different colored lights to distinguish the instruction status, so as to complete the two-way confirmation interaction between human and machine. The permission control module has a built-in hierarchical permission control system, which works in conjunction with the AI ​​intelligent verification module and the security limit judgment module to classify operation permissions based on patient identity, physical condition and nursing level, and to distinguish between ordinary daily operation permissions, restricted operation permissions and prohibited operation permissions. The safety limit determination module pre-stores the safety thresholds and limit parameters of each action of the nursing bed, including the lifting height limit, backrest tilting angle limit, leg lifting angle limit, side tilting angle limit, and action speed threshold. It is used to perform safety verification on the confirmed control commands and determine whether the command operation exceeds the safety range of the equipment and whether it is suitable for the patient's physical tolerance. The bed drive control module is connected to the safety limit determination module and the power actuator of the intelligent nursing bed, respectively. It is used to receive legal and safe commands that have passed verification and drive the nursing bed to complete corresponding actions such as lifting, turning, leg raising, side movement, and lying flat.

[0007] Furthermore, the recognition process of the big data speech recognition module includes: collecting original spoken speech → extracting speech spectrum features and pronunciation features → matching non-standard pronunciation samples in the big data speech database → correcting colloquial ambiguities → outputting standardized control commands, which can recognize various patient speech commands such as omitted sentences, inverted sentences, colloquial particles, and unclear pronunciation.

[0008] Furthermore, the confirmation mechanism of the interactive feedback module is as follows: for clear and valid instructions, the system will announce "The operation XX is about to be performed, do you confirm?"; for ambiguous instructions, the system will announce "The instruction was not clearly identified, is operation XX or operation XX required?" to guide the patient to confirm again; for invalid or non-compliant instructions, the system will directly announce "This operation cannot be performed" and terminate the instruction transmission.

[0009] Furthermore, the access control module has tiered access permissions, including: Level 1 general access, which is applicable to all patients and allows basic safe operations such as lying flat, slight lifting and lowering, and slight leg raising; Level 2 restricted access, which is only available after authorization by medical staff and allows operations such as large-angle adjustment and body position rotation; and Level 3 prohibited access, which permanently locks dangerous and excessive operations, including excessive height lifting and lowering and extreme side rolling.

[0010] Furthermore, the safety limit determination module is linked in real time with the nursing bed pressure sensor and posture sensor to dynamically collect patient position data and bed operation data. If the instruction operation will cause the bed to operate beyond the threshold, the patient's position to become unbalanced, or the limbs to be compressed, the instruction will be directly intercepted and a safety warning will be triggered.

[0011] Furthermore, the system has a built-in memory learning unit that can record patients' spoken pronunciation habits and commonly used commands in real time, continuously update the local voice sample library, and gradually adapt to the non-standard pronunciation characteristics of individual patients to improve long-term recognition accuracy.

[0012] Compared with the prior art, the present invention has the following significant advantages: This invention features a multi-dimensional big data voice library, containing a massive collection of non-standard Mandarin, colloquial, and dialectal Mandarin voice samples. Breaking through the limitations of traditional voice systems that only support standard Mandarin and fixed phrases, it can accurately recognize patient voice commands with non-standard pronunciation, colloquial expressions, and irregular sentence structures. It is suitable for various groups, including elderly patients, patients from dialect regions, and patients with speech disorders, significantly improving the adaptability and error tolerance of voice interaction. A two-way confirmation feedback mechanism is established, using AI intelligent verification to filter ambiguous, vague, and invalid commands, proactively repeating and confirming the operation requirements to the patient. This avoids mis-execution problems caused by voice recognition errors or patients' casual expressions, achieving a precise closed-loop command interaction and improving operational accuracy. It integrates an AI-based hierarchical permission control and a dual protection system of safety limits. By dividing permissions to control the operable range in different scenarios, and using safety thresholds to intercept excessive and dangerous operations, it provides dual protection from both operational permission and equipment operational safety dimensions. This effectively avoids secondary injuries such as falls, limb compression, and postural injuries caused by patient misoperation or exceeding limits, greatly improving the safety of using the intelligent nursing bed. It has the ability to learn and optimize itself, and can continuously iterate and optimize the recognition model to adapt to the patient's unique pronunciation habits. With long-term use, the recognition accuracy gradually improves. It does not require tedious manual debugging, adapts to various nursing scenarios, and takes into account convenience, intelligence and safety, greatly improving the practical value and adaptability of the smart nursing bed. Attached Figure Description

[0013] Figure 1 This is a block diagram of the overall system architecture of the present invention; Figure 2 This is a schematic diagram of the speech recognition and security control process of the present invention.

[0014] The module includes: 1. Voice acquisition module; 2. Big data voice recognition module; 3. AI intelligent verification module; 4. Interactive feedback module; 5. Access control module; 6. Security limit determination module; and 7. Bed drive control module. Detailed Implementation

[0015] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0016] This invention discloses a voice control system for an intelligent nursing bed, such as... Figure 1 As shown, the system consists of a voice acquisition module 1, a big data voice recognition module 2, an AI intelligent verification module 3, an interactive feedback module 4, an access control module 5, a security limit determination module 6, and a bed drive control module 7. The modules work together to complete the entire process of voice acquisition, intelligent recognition, verification and confirmation, security determination, and equipment drive control.

[0017] The specific operation process is as follows: Step 1, Voice Acquisition: The voice acquisition module 1 uses a dual-microphone noise reduction array to acquire the patient's spoken voice commands in real time. It automatically filters out ambient noise in the ward, air conditioning noise, and equipment operating noise, purifying the original voice signal and avoiding interference from noise to improve recognition accuracy. Patients can use any spoken expression and non-standard Mandarin pronunciation to give commands without having to follow a fixed standard script.

[0018] Step 2, Non-standard speech intelligent recognition: The big data speech recognition module 2 calls up the built-in cloud big data speech library, extracts the spectrum features and pronunciation features of the original speech signal, and performs semantic correction, fuzzy matching and command conversion for problems such as non-standard patient pronunciation, inverted sentence structure, excessive spoken auxiliary words, and ambiguous expression. It accurately converts non-standard spoken speech into standardized control commands that the nursing bed can recognize, and completes non-discriminatory speech recognition.

[0019] Step 3: AI Intelligent Verification and Two-Way Feedback: The AI ​​intelligent verification module 3 receives standardized control commands and performs semantic verification based on the patient's historical operating habits and the current nursing scenario, distinguishing between clear, ambiguous, invalid, and dangerous commands. For clear commands, the commands are transmitted to the interactive feedback module 4, which replies the commands via voice and requests the patient's confirmation. For ambiguous commands, alternative operation options are proactively pushed to guide the patient's confirmation. Invalid commands are directly rejected, achieving two-way interactive confirmation between humans and machines and eliminating recognition errors.

[0020] Step 4, Permission Level Verification: The permission control module 5 verifies the permissions of the confirmed instructions based on the patient's identity information, nursing level, and medical authorization status. It determines whether the current instruction is within the scope of the patient's operation, and blocks restricted or prohibited operations without permission, thus avoiding the risk of unauthorized operations from the source.

[0021] Step 5, Safety Limit Determination: The safety limit determination module 6 retrieves the pre-stored safe operating thresholds for the nursing bed and, in conjunction with real-time collected bed posture data and patient position data, verifies whether the bed movements corresponding to the command exceed the safety limits and whether they will cause risks such as pressure on the patient or a fall from the bed. If a safety hazard exists, the command is immediately intercepted and a voice warning is triggered; if it is safe and compliant, the command is allowed.

[0022] Step 6, Equipment Drive Execution: The bed drive control module 7 receives the finally verified safe and legal instructions and drives the nursing bed power actuator to complete the corresponding lifting, backrest adjustment, leg raising, body position rotation and other actions, and completes the intelligent voice control operation.

[0023] Meanwhile, the system has a built-in memory learning unit that continuously collects patients' pronunciation features and common spoken commands during daily use, iteratively updates the local speech sample library, and continuously optimizes the exclusive recognition model. With long-term use, the recognition accuracy continues to improve, and the adaptability is stronger.

[0024] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A voice control system for an intelligent nursing bed, characterized in that: It includes a voice acquisition module, a big data voice recognition module, an AI intelligent verification module, an interactive feedback module, an access control module, a security limit determination module, and a bed drive control module; The voice acquisition module is used to acquire patients' spoken voice commands in real time, filter environmental noise, and output pure original voice data. The big data speech recognition module is electrically connected to the speech acquisition module and the AI ​​intelligent verification module respectively. The big data speech recognition module has a built-in cloud-linked multi-dimensional big data speech library. The big data speech library includes multi-regional non-standard Mandarin pronunciation, colloquial and casual speech, and dialect Mandarin variant data, which are used to extract features, semantically correct, and perform fuzzy matching on non-standard and colloquial speech commands, and convert them into standardized equipment control commands. The AI ​​intelligent verification module is used to receive standardized control commands, combine semantic logic with the patient's operating habits to complete intelligent verification, filter ambiguous, vague and invalid commands and generate feedback messages. The interactive feedback module has a built-in voice broadcasting unit and light prompting unit, which are used to repeat the instructions to be executed to the patient and initiate secondary confirmation, so as to realize two-way human-computer interaction. The permission control module has a built-in hierarchical permission control system, which is used to classify normal operation, restricted operation, and prohibited operation permissions based on patient identity and nursing level, and to complete instruction permission verification. The safety limit determination module pre-stores the safety threshold and limit parameters of the nursing bed's actions, which are used to verify the safety of command operations and intercept excessive or dangerous operations. The bed drive control module is used to receive compliant and safe instructions that have passed verification and drive the intelligent nursing bed to complete the corresponding actions.

2. The intelligent nursing bed voice control system according to claim 1, characterized in that: The voice acquisition module adopts a dual-microphone noise reduction array structure, which is located on both sides of the head of the smart nursing bed. It can filter environmental noise, breathing noise, and equipment operation noise, and is suitable for the complex noise environment of the ward.

3. The intelligent nursing bed voice control system according to claim 1, characterized in that: The big data speech recognition module can recognize non-standard spoken commands with omitted sentences, inverted sentences, mixed spoken particles, and unclear pronunciation, and is suitable for the speech interaction needs of patients with dialect Mandarin and non-standard Mandarin.

4. The intelligent nursing bed voice control system according to claim 1, characterized in that: The confirmation mechanism of the interactive feedback module includes three interaction modes: confirmation of clear and valid instructions, confirmation of alternative instructions for ambiguous or unclear instructions, and immediate rejection and broadcast of invalid or illegal instructions.

5. The intelligent nursing bed voice control system according to claim 1, characterized in that: The access control module has three levels of permissions: Level 1 general permissions, Level 2 restricted permissions, and Level 3 prohibited permissions. Level 2 restricted permissions require authorization from medical staff to be granted, while Level 3 prohibited permissions permanently lock out dangerous over-limit operations.

6. The intelligent nursing bed voice control system according to claim 1, characterized in that: The safety limit determination module is linked in real time with the pressure sensor and posture sensor of the nursing bed to dynamically collect patient position and bed operation data, and determine whether there is a risk of positional imbalance, limb pressure, or equipment exceeding the limit during operation.

7. The intelligent nursing bed voice control system according to claim 1, characterized in that: The system has a built-in memory learning unit that can record patients' spoken pronunciation habits and commonly used commands, iteratively update the local speech sample library, and adaptively optimize recognition accuracy.

Images