Internet of Things forklift management system and method

Abstract

The invention relates to an internet-of-things forklift management system and method, the system comprises gateway equipment, an ignition switch, mobile phone software, cockpit acquisition equipment, blind area detection equipment, a rear-end system and alarm equipment, and the gateway equipment, the ignition switch, the cockpit acquisition equipment, the blind area detection equipment and the alarm equipment are all installed on a forklift. The ignition switch, the cockpit acquisition equipment, the blind area detection equipment and the alarm equipment are connected with the gateway equipment, the gateway equipment is in communication connection with the rear-end system, and the rear-end system is in communication connection with the mobile phone software; face identity verification is completed by adopting mobile phone software and a back-end system, so that the implementation cost of the whole solution is greatly reduced. And meanwhile, an indoor and outdoor fusion positioning function is integrated on the gateway equipment side, so that the forklift can accurately lock the position of the forklift no matter the forklift is located in the open air or in a field area, and the supervision capability of an enterprise on the forklift is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of forklift management, and in particular relates to an Internet of Things forklift management system and method. Background Art

[0002] As an indispensable handling equipment in manufacturing, warehousing and logistics, intelligent supervision of forklifts has become an industry development trend. In particular, in recent years, frequent forklift safety accidents have caused casualties and property losses, which have brought profound lessons to employers and emergency management units. Therefore, the implementation of intelligent supervision of forklifts is also imminent.

[0003] At present, there are some relatively mature solutions for the intelligent supervision of forklifts, but due to several difficult problems, the existing supervision solutions are not easy to promote. First, the existing solutions are priced high, which makes the burden on enterprises heavy. Most of the current solutions are given to AI edge computing units for the development of overall solutions, which makes the cost of the solutions high. In addition, since multiple AI algorithms need to be run on the AI ​​edge computing unit, and multiple algorithms are not run at the same time, it has caused a certain amount of waste. Second, the existing solutions have low hazard identification accuracy, which makes the system ineffective after it goes online. For solutions with higher accuracy, the cost must be extremely high, and if the cost of the solution is controlled blindly, it will inevitably lead to low identification accuracy of the existing solutions. Since enterprises cannot choose a more reasonable balance between accuracy and cost, the system did not play the expected role after it went online. Third, there are blind spots for forklift supervision in the field, which makes it impossible to effectively supervise forklifts. The current solutions all support GPS / Beidou positioning functions, but because the company has large indoor fields, once the forklift enters the room, the positioning signal is completely missing. Fourth, the existing solutions have poor scalability, making it difficult for users to customize functions.

[0004] Based on the above shortcomings of the existing solutions and the necessity of intelligent forklift supervision, there is an urgent need for an Internet of Things forklift management system and method to promote the promotion of intelligent supervision in this field. Summary of the invention

[0005] The present invention provides an Internet of Things forklift management system and method to achieve low-cost intelligent management of forklifts, thereby solving the technical problems mentioned in the background technology.

[0006] A technical solution of the present invention is as follows: An Internet of Things forklift management system, including: a gateway device, an ignition switch, a mobile phone software, a cockpit collection device, a blind spot detection device, a back-end system and an alarm device, wherein the gateway device, the ignition switch, the cockpit collection device, the blind spot detection device and the alarm device are all installed on a forklift.

[0007] The ignition switch, cockpit acquisition equipment, blind spot detection equipment, and alarm equipment are connected to the gateway device, the gateway device is communicatively connected to the backend system, and the backend system is communicatively connected to the mobile phone software; The mobile phone software can obtain the driver's facial information, and the back-end system can perform identity recognition based on the facial information. When the identity recognition is passed, an ignition signal is generated, and the gateway device can control the ignition switch to start the forklift based on the ignition signal.

[0008] The gateway device can locate the position of the forklift through wireless communication signals and satellite positioning signals to generate positioning information.

[0009] The cockpit acquisition device can capture the driver during the driving of the forklift to generate a driving photo of the driver, and the back-end system can detect the driving behavior based on the driving photo of the driver.

[0010] The blind spot detection device can detect obstacles in the driver's blind spot during the driving of the forklift. When an obstacle is detected, an alarm signal is generated. The gateway device can control the alarm device to sound an alarm according to the alarm signal.

[0011] Furthermore, the gateway device includes a processing module, a satellite positioning module and a wireless communication module. The processing module is connected to the satellite positioning module and the wireless communication module respectively. The processing module communicates with the back-end system through the wireless communication module and the wireless communication module can receive wireless communication signals indoors. The satellite positioning module can obtain satellite positioning signals. The processing module can determine whether the forklift is in an indoor scene or an outdoor scene based on the strength of the satellite positioning signal. When the forklift is in the outdoor scene, the processing module can determine the outdoor positioning information of the forklift based on the satellite positioning signal; when the forklift is in the indoor scene, the processing module can determine the indoor positioning information of the forklift based on the wireless communication signal.

[0012] Furthermore, the satellite positioning module includes a GNSS module, and the GNSS module is connected to the processor via a UART interface.

[0013] Furthermore, the communication module includes a 4G module, a WiFi module and a Bluetooth module.

[0014] Furthermore, the gateway device is provided with a power module, and the power module supplies power to the gateway device.

[0015] Furthermore, the back-end system can detect driving behavior based on the driver's driving photos, including: the back-end system detects uncivilized driving behavior through an image recognition algorithm, and when uncivilized driving behavior occurs, the driver's driving photos are retained and a reminder alarm signal is generated, and the gateway device can control the alarm device to sound an alarm based on the reminder alarm signal.

[0016] Furthermore, a vehicle-specific QR code is provided on the gateway device. The driver scans the vehicle-specific QR code through a mobile phone software, and the background system determines the vehicle that the driver currently needs to start based on the vehicle-specific QR code.

[0017] Further, the blind spot detection device comprises: a detection processor, an image sensor and a communication module, the image sensor and the communication module are both connected to the detection processor, and the detection processor is connected to the gateway device through the communication module; The image sensor can obtain a blind spot image during the travel of the forklift, and the detection processor performs human figure detection on the blind spot image.

[0018] Furthermore, the human detection algorithm of the detection processor adopts weight pruning and conduction pruning as algorithm pruning strategies, adopts a quantization strategy of quantization-aware training, and adopts a distillation strategy for model training.

[0019] Another technical solution of the present invention is as follows: an Internet of Things forklift management method, using any one of the Internet of Things forklift management systems described in the above claims, comprising: The driver obtains the driver's facial information through the mobile phone software, and the back-end system performs identity recognition. When the identity recognition is passed, the back-end system generates an ignition signal, and the gateway device can control the ignition switch to start the forklift according to the ignition signal; When the driver is working, the cockpit acquisition device can capture the driver during the driving of the forklift and generate a driving photo of the driver. The back-end system can detect the driving behavior based on the driving photo of the driver; The blind spot detection device can detect obstacles in the driver's blind spot during the driving of the forklift, and when an obstacle is detected, an alarm signal is generated, and the gateway device can control the alarm device to alarm according to the alarm signal; The driver completes the work and turns off the engine by turning on the ignition switch.

[0020] Beneficial effects of the present invention: Low cost: By using mobile phone software to complete facial identity verification, removing the traditional AI edge computing unit, using blind spot detection equipment with built-in AI algorithms, and completing driver behavior recognition in the back-end system, the overall solution cost is reduced by more than 40% compared to traditional methods.

[0021] High performance: By adopting blind spot detection equipment with built-in AI algorithm, the blind spot recognition accuracy can be improved to 99%; by integrating the fusion positioning function on the gateway device side, the problem of GPS / Beidou positioning being unable to locate indoors can be completely solved.

[0022] High reliability: The inventions described above all use relatively mature equipment and technologies, and do not involve any precision instruments, high-end equipment, or the like. Therefore, by organically and efficiently combining these equipment and technologies, not only can the overall solution be guaranteed to be stable and reliable, but it can also have good scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a structural block diagram of the present invention.

[0024] Figure 2 It is a schematic diagram of the structure of the gateway device in the present invention.

[0025] Figure 3 It is a structural schematic diagram of the blind spot detection device in the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. The described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0027] In one technical solution of the present invention, Figure 1 It is a structural block diagram provided according to the specific structure of an Internet of Things forklift management system of the present invention, such as Figure 1 As shown, the present invention includes: a gateway device 1, an ignition switch 4, a mobile phone software 3, a cockpit collection device 5, a blind spot detection device 6, a back-end system 2 and an alarm device 7, and the gateway device 1, the ignition switch 4, the cockpit collection device 5, the blind spot detection device 6 and the alarm device 7 are all installed on a forklift.

[0028] The gateway device 1 realizes the control, processing and network connection of the forklift management system. The ignition switch 4, the cockpit acquisition device 5, the blind spot detection device 6, and the alarm device 7 are connected to the gateway device 1, the gateway device 1 is connected to the backend system 2, and the backend system 2 is connected to the mobile phone software 3.

[0029] The back-end system 2 may be a server, in which platform management software is run, thereby achieving comprehensive management of the forklift.

[0030] The mobile phone software 3 can obtain the driver's facial information, and the back-end system 2 can perform identity recognition based on the facial information. When the identity recognition is passed, an ignition signal is generated, and the gateway device 1 can control the ignition switch 4 to start the forklift according to the ignition signal.

[0031] Among them, the facial information is a facial photo, and the mobile phone software 3 is a separately developed mobile phone software 3 that can call the mobile phone camera to take a facial photo of the driver and send it to the back-end system 2, which performs facial recognition.

[0032] The gateway device 1 can locate the position of the forklift through wireless communication signals and satellite positioning signals to generate positioning information.

[0033] The cockpit acquisition device 5 can capture the driver during the driving of the forklift to generate a driving photo of the driver, and the back-end system 2 can detect the driving behavior based on the driving photo of the driver; The blind spot detection device 6 can detect obstacles in the driver's blind spot during the driving of the forklift. When an obstacle is detected, an alarm signal is generated. The gateway device 1 can control the alarm device 7 to sound an alarm according to the alarm signal.

[0034] Specific: Mobile phone software 3: It is mainly used to scan the vehicle QR code to complete the facial identity verification function. At the same time, it can use the mobile phone’s own positioning module to complete the position comparison of people and vehicles to avoid the separation of people and vehicles.

[0035] Back-end system 2: Mainly used for back-end monitoring by enterprises and competent authorities, and also completes functions such as identity information verification and comparison, data analysis, and remote supervision.

[0036] Gateway device 1: Mainly used to connect the backend system 2 with the local IoT devices, supporting 4G communication, GNSS positioning, indoor and outdoor integrated positioning, outdoor GPS / Beidou positioning, indoor Bluetooth / WiFi positioning, and can confirm whether the current vehicle is located outdoors or indoors based on the strength of the GPS / Beidou positioning signal, and perform seamless logic switching based on the received outdoor positioning signal and indoor positioning signal. At the same time, it can be compatible with video recorders NVR, screen displays and other functions according to storage requirements.

[0037] Ignition switch 4: Mainly used to control vehicle start. Specifically, it is ACC ignition switch 4 Cockpit acquisition equipment 5: mainly used for video monitoring of the driver, specifically an AHD camera.

[0038] Blind spot detection device 6: It is mainly used for safety warning of blind spots around vehicles. It has built-in AI algorithm and can independently complete monitoring, identification, triggering alarm and other functions. It can be a BSD camera.

[0039] Alarm device 7: mainly used for broadcasting alarm information.

[0040] In one embodiment of the present technical solution: Figure 2 As shown, the gateway device 1 includes a processing module 11, a satellite positioning module 12 and a wireless communication module 13. The processing module 11 is connected to the satellite positioning module 12, the wireless communication module 13 and the wireless communication module 13 respectively. The processing module 11 communicates with the back-end system 2 through the wireless communication module 13 and receives indoor wireless communication signals. The satellite positioning module 12 can obtain satellite positioning signals. The processing module 11 can determine whether the forklift is in an indoor scene or an outdoor scene according to the strength of the satellite positioning signal. When the forklift is in the outdoor scene, the processing module 11 can determine the outdoor positioning information of the forklift according to the satellite positioning signal; when the forklift is in the indoor scene, the processing module 11 can determine the indoor positioning information of the forklift according to the wireless communication signal.

[0041] Specifically, the positioning information of the forklift includes indoor positioning information and outdoor positioning information. The processing module 11 analyzes and obtains the signal strength of the current satellite positioning module 12. When it is lower than the set threshold, it is determined that the forklift is in an indoor scene. The indoor scene is positioned by relying on the signal strength of the wireless communication module 13, specifically the Bluetooth signal or the WIFI signal. It should be noted that the Bluetooth signal or WIFI signal positioning technology is well known to technicians in this field, so it will not be repeated here.

[0042] The satellite positioning module 12 includes a GNSS module, which is connected to the processor via a UART interface. The wireless communication module 13 includes a 4G module, a WiFi module and a Bluetooth module. The gateway device 1 is provided with a power module 14, which supplies power to the gateway device 1. The processing module 11 may specifically adopt an MCU chip.

[0043] The gateway device 1 integrates the processing module 11, the 4G module including the 4G module, the WiFi module and the Bluetooth module, the power management, and the GNSS module, and reserves communication interfaces with the ACC ignition switch 4, the AHD camera, the BSD camera, the sound and light alarm and other devices, so as to complete the on-site deployment and closed-loop work of the entire system.

[0044] The 4G module is mainly used for communication between the local and background systems. The 4G module is connected to the main processor via UART. The main processor will drive the ACC, AHD cameras and other devices to perform actions such as switching on and off, capturing, etc. through the I / O interface according to the 4G remote instructions.

[0045] The power module 14 is mainly used to provide stable support of different high and low power supplies for each core component of the intelligent gateway to ensure that all components work normally.

[0046] The GNSS module supports Beidou and GPS multi-mode positioning. The main processor collects GNSS positioning data regularly through the UART interface, and sends the data to the background through the 4G module to support background completion, vehicle positioning, electronic fence and other location services.

[0047] In one embodiment of the present technical solution: the back-end system 2 can perform driving behavior detection based on the driver's driving photos, including: the back-end system 2 detects uncivilized driving behavior through an image recognition algorithm, and when uncivilized driving behavior occurs, the driver's driving photos are retained and a reminder alarm signal is generated, and the gateway device 1 can control the alarm device 7 to sound an alarm based on the reminder alarm signal.

[0048] Among them, uncivilized driving behaviors specifically include: not wearing a helmet, not wearing a seat belt, smoking, answering or making phone calls, etc. At the same time, the back-end system 2 can also perform off-duty detection. The detection results can be archived in the back-end on the one hand, and local alarm reminders can be issued through the gateway device 1 on the other hand.

[0049] In one embodiment of the present technical solution: the gateway device 1 is provided with a vehicle-specific QR code, and the driver scans the vehicle-specific QR code through the mobile phone software 3, and the backend system determines the vehicle that the driver currently needs to start based on the vehicle-specific QR code. At the same time, the backend system 2 obtains the driver's photo taken by the mobile phone software 3 and compares it with the photo of the driver who needs to start the vehicle, so as to determine whether the current driver corresponds to the vehicle. If so, the verification result is sent to the gateway device 1 and the mobile phone software 3 through the 4G network. The mobile phone software 3 will display the "verification passed" message, and the gateway device 1 will simultaneously trigger the ignition switch 4 to turn on; otherwise, the mobile phone software 3 will display the "verification failed" message, and the gateway device 1 will not trigger the ignition switch 4.

[0050] In one embodiment of the present technical solution: the blind spot detection device 6 includes: a detection processor 61, an image sensor 62 and a communication module 63, the image sensor 62 and the communication module 63 are both connected to the detection processor 61, the detection processor 61 is connected to the gateway device 1 through the communication module, the image sensor 62 can obtain the blind spot image during the driving process of the forklift, and the detection processor 61 performs human shape detection on the blind spot image.

[0051] The human detection algorithm of the detection processor 61 adopts weight pruning and conduction pruning as algorithm pruning strategies, adopts a quantization strategy of quantized perceptual training, and adopts a distillation strategy for model training.

[0052] Specific: The schematic diagram of the BSD blind spot detection camera with AI recognition function is as follows: Figure 3 As shown. Different from ordinary cameras, the BSD camera of the present invention realizes the function of the camera body completing AI detection by means of a built-in image AI recognition unit. In addition, the camera ISP main processor completes interface communication with WiFi / RJ45 communication, data storage, image sensor (62) and other parts through RMII, SDIO, MIPI and other methods.

[0053] The image AI recognition unit can integrate different types of AI algorithms according to the different functions to be realized by the camera, so as to realize detection cameras with different AI algorithm functions. In the present invention, the image AI recognition unit will integrate a human detection algorithm to realize real-time and on-site detection of pedestrians.

[0054] In order to reduce the cost of equipment, the present invention generally uses a processor with lower computing power to support the algorithm work that meets the detection requirements, so it is necessary to prune and optimize the ordinary AI algorithm. The AI ​​algorithm in the present invention is a human detection algorithm. In the implementation of the algorithm, the algorithm will use a combination of lightweight basic models, pruning, quantized perception training, and knowledge distillation to improve the accuracy and running efficiency of the human detection algorithm. The optimal path, so as to maximize the use of chip performance according to the hardware characteristics of the selected image recognition unit. Specifically, the present invention uses a lightweight version of YoLov8-nano with a better structural design as the basic model to reduce the difficulty of subsequent optimization. Weight pruning and conduction pruning are used as algorithm pruning strategies, in which weight pruning will reduce redundant calculations by sparse weight distribution; while channel pruning will retain the key channels in the convolution (the channels that can most affect the detection head) and delete the channels that have less impact on human detection; after the completion of the two methods, the model will also be slightly fine-tuned to improve the accuracy of the algorithm. In addition, in order to enable the low-computing-power image AI recognition unit to achieve computational acceleration while maintaining high algorithm accuracy, the present invention will adopt a quantization strategy of quantization-aware training to quantize the weights to INT8. At the same time, a distillation strategy is used to complete the training of a large model with high accuracy to guide the lightweight small model, that is, by extracting the feature map of the large model, forcing the small model to learn its feature distribution, and performing multi-task distillation for classification probability, human-shaped box regression target and intermediate layer features, so that the small model can also achieve the accuracy of the large model.

[0055] In another technical solution of the present invention, the present technical solution provides an Internet of Things forklift management method, which specifically adopts any of the above-mentioned Internet of Things forklift management systems, including: The driver obtains the driver's facial information through the mobile phone software 3, and the back-end system 2 performs identity recognition. When the identity recognition is passed, the back-end system 2 generates an ignition signal. The gateway device 1 can control the ignition switch 4 to start the forklift according to the ignition signal.

[0056] The driver first powers on the forklift and scans the vehicle-specific QR code through the mobile phone software 3 for facial identity comparison. The QR code is uniformly provided by the back-end system 2.

[0057] The mobile phone software 3 can be connected to the back-end system 2. After receiving the driver's facial identity comparison message, the back-end system 2 performs identity verification and sends the verification result to the gateway device 1 and the mobile phone software 3 through the 4G network.

[0058] If the verification result of the back-end system 2 is passed, the mobile phone software 3 will display the message "Verification passed", and the gateway device 1 will simultaneously trigger the ACC ignition switch 4 to turn on; otherwise, the mobile phone software 3 will display the message "Verification failed", and the gateway device 1 will not trigger the ACC ignition switch 4.

[0059] When the driver is working, the cockpit acquisition device 5 can capture the driver during the driving of the forklift and generate a driving photo of the driver. The back-end system 2 can detect the driving behavior based on the driver's driving photo.

[0060] The driver drives the vehicle for normal operation, and the AHD high-definition camera regularly captures photos of the driver's driving operation and transmits them to the back-end system 2 through the intelligent gateway. The back-end system 2 performs behavior recognition on the pictures captured by the AHD camera, and can complete safety helmet detection, seat belt detection, smoking detection, phone call detection, and off-duty detection according to regulatory requirements. The detection results are archived in the back-end on the one hand, and local alarm reminders can be issued through the gateway device 1 on the other hand.

[0061] The blind spot detection device 6 can detect obstacles in the driver's blind spot during the driving of the forklift. When an obstacle is detected, an alarm signal is generated. The gateway device 1 can control the alarm device 7 to sound an alarm according to the alarm signal.

[0062] The BSD blind spot camera determines in real time whether there is any abnormality around the vehicle. If so, it drives the alarm through the gateway device 1 to give a real-time alarm, otherwise it continues to work.

[0063] The driver completes the operation and turns off the engine by turning on the ignition switch 4. The driver completes the operation task, turns off the power, and the system operation ends.

[0064] The present invention is a low-cost solution, that is, by using mobile phone software 3 and back-end system 2 to complete face identity verification, removing the traditional AI edge computing unit, using BSD camera with built-in AI algorithm, and completing the driver behavior recognition algorithm at the back end, the implementation cost of the overall solution is greatly reduced. At the same time, the indoor and outdoor fusion positioning function is integrated on the gateway device 1 side, so that the forklift can accurately lock the vehicle position whether it is located outdoors in the open air or indoors in the field area, thereby improving the company's supervision ability over the vehicle.

[0065] Finally, it should be noted that the above specific implementation methods are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. An Internet of Things forklift management system, characterized in that: include: A gateway device (1), an ignition switch (4), a mobile phone software (3), a cockpit data acquisition device (5), a blind spot detection device (6), a back-end system (2) and an alarm device (7), wherein the gateway device (1), the ignition switch (4), the cockpit data acquisition device (5), the blind spot detection device (6) and the alarm device (7) are all installed on a forklift; The ignition switch (4), cockpit data acquisition device (5), blind spot detection device (6), and alarm device (7) are connected to the gateway device (1), the gateway device (1) is communicatively connected to the back-end system (2), and the back-end system (2) is communicatively connected to the mobile phone software (3); The mobile phone software (3) can obtain the driver's facial information, the back-end system (2) can perform identity recognition based on the facial information, and when the identity recognition is passed, an ignition signal is generated, and the gateway device (1) can control the ignition switch (4) to start the forklift according to the ignition signal; The gateway device (1) is capable of locating the position of the forklift through wireless communication signals and satellite positioning signals, and generating positioning information; The cockpit acquisition device (5) can capture the driver during the driving of the forklift to generate a driving photo of the driver, and the back-end system (2) can detect the driving behavior based on the driving photo of the driver; The blind spot detection device (6) can detect obstacles in the driver's blind spot during the driving of the forklift, and when an obstacle is detected, an alarm signal is generated. The gateway device (1) can control the alarm device (7) to sound an alarm according to the alarm signal.

2. The Internet of Things forklift management system according to claim 1, characterized in that: The gateway device (1) comprises a processing module (11), a satellite positioning module (12) and a wireless communication module (13); the processing module (11) is connected to the satellite positioning module (12) and the wireless communication module (13) respectively; the processing module (11) communicates with the back-end system (2) via the wireless communication module (13) and the wireless communication module (13) is capable of receiving wireless communication signals indoors; the satellite positioning module (12) is capable of acquiring satellite positioning signals; the processing module (11) is capable of determining whether the forklift is in an indoor scene or an outdoor scene according to the strength of the satellite positioning signal; when the forklift is in an outdoor scene, the processing module (11) is capable of determining the outdoor positioning information of the forklift according to the satellite positioning signal; when the forklift is in an indoor scene, the processing module (11) is capable of determining the indoor positioning information of the forklift according to the wireless communication signal.

3. The Internet of Things forklift management system according to claim 2, characterized in that: The satellite positioning module (12) comprises a GNSS module, and the GNSS module is connected to the processor via a UART interface.

4. The Internet of Things forklift management system according to claim 2, characterized in that: The wireless communication module (13) comprises a 4G module, a WiFi module and a Bluetooth module.

5. The Internet of Things forklift management system according to claim 2, characterized in that: The gateway device (1) is provided with a power module (14), and the power module (14) supplies power to the gateway device (1).

6. The Internet of Things forklift management system according to claim 1, characterized in that: The back-end system (2) is capable of detecting driving behavior based on the driver's driving photo, including: the back-end system (2) detects uncivilized driving behavior through an image recognition algorithm, retains the driver's driving photo when uncivilized driving behavior occurs, and generates a warning alarm signal; the gateway device (1) is capable of controlling the alarm device (7) to sound an alarm based on the warning alarm signal.

7. The Internet of Things forklift management system according to claim 1, characterized in that: The gateway device (1) is provided with a vehicle-specific QR code. The driver scans the vehicle-specific QR code through the mobile phone software (3), and the backend system determines the vehicle that the driver currently needs to start based on the vehicle-specific QR code.

8. The Internet of Things forklift management system according to claim 1, characterized in that: The blind spot detection device (6) comprises: a detection processor (61), an image sensor (62) and a communication module (63), wherein the image sensor (62) and the communication module (63) are both connected to the detection processor (61), and the detection processor (61) is connected to the gateway device (1) via the communication module (63); The image sensor (62) is capable of acquiring a blind spot image during the travel of the forklift, and the detection processor (61) performs human figure detection on the blind spot image.

9. The Internet of Things forklift management system according to claim 1, characterized in that: The human detection algorithm of the detection processor (61) adopts weight pruning and conduction pruning as algorithm pruning strategies, adopts a quantization strategy of quantization perception training, and adopts a distillation strategy for model training.

10. An Internet of Things forklift management method, characterized in that: The Internet of Things forklift management system according to any one of claims 1 to 9 comprises: The driver obtains the driver's facial information through the mobile phone software (3), and the back-end system (2) performs identity recognition. When the identity recognition is passed, the back-end system (2) generates an ignition signal, and the gateway device (1) can control the ignition switch (4) to start the forklift according to the ignition signal; When the driver is operating, the cockpit acquisition device (5) can capture the driver during the driving of the forklift to generate a driving photo of the driver, and the back-end system (2) can detect the driving behavior based on the driving photo of the driver; The blind spot detection device (6) is capable of detecting obstacles in the driver's blind spot during the driving of the forklift, and when an obstacle is detected, an alarm signal is generated, and the gateway device (1) is capable of controlling the alarm device (7) to sound an alarm according to the alarm signal; The driver completes the operation and turns off the engine by turning on the ignition switch (4).

Images