Traffic radar and video integrated machine adaptive tuning system and method
Through the adaptive tuning system, the frequency and switching data of traffic radar and video all-in-one machine are dynamically adjusted, which solves the problem of working instability caused by voltage interference, and achieves the effect of improving equipment stability and reducing costs.
Patent Information
- Application Number
- CN202411105105.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-08-13
AI Technical Summary
Traffic radar and video all-in-ones are susceptible to voltage interference fluctuations when installed on municipal roads, resulting in unstable work. The existing technology solves the problem by adding multi-stage filtering and voltage stabilization modules, but increases costs and makes it difficult to deploy on a large scale.
Adaptive tuning system is adopted, through power modules, voltage conversion modules, acquisition modules, equation building modules and adaptive tuning modules, real-time acquisition and analysis of branch voltage, frequency and switching data, dynamically adjust frequency and switching data to reduce voltage Comprehensive fluctuations to improve equipment stability.
There is no need to add additional hardware equipment, which effectively reduces the voltage fluctuations of traffic radar and video all-in-one machines, improves the working stability of the equipment, reduces the cost of voltage stabilization, and is suitable for large-scale deployment.
Smart Images

Figure CN118642095B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of radar and video all-in-one machines, and in particular to a traffic radar and video all-in-one machine adaptive tuning system and method. Background Art
[0002] Traffic radar can be used in applications such as traffic signal control, traffic flow monitoring, vehicle counting and speed detection. Video surveillance system is a device that uses cameras and video recording equipment to monitor and record traffic scenes. It can monitor traffic conditions in real time, record traffic accidents and violations, and provide evidence for investigation and proof. Traffic radar and video all-in-one integrates traffic radar and video surveillance functions, and can provide radar detection and video surveillance functions at the same time. It can monitor traffic flow, vehicle speed and direction in real time, and record traffic scenes through video surveillance. Traffic radar and video all-in-one is a device that integrates radar and video surveillance functions. It is used to detect and track traffic flow using radar technology, thereby realizing traffic monitoring and safety management. It can measure the speed, distance and direction of vehicles by sending and receiving radio waves. This integrated device can provide more comprehensive and accurate traffic information, helping traffic management departments to carry out traffic planning, traffic flow management and traffic safety monitoring.
[0003] Traffic radar and video all-in-one devices usually have high-precision radar detection and video monitoring functions, and can adapt to various traffic scenes and environments. They are usually installed in traffic intersections, highways, parking lots and other places for traffic management and safety monitoring.
[0004] At present, the power supply method used by traffic radar and video all-in-one is to convert the input AC 220V through AC-DC, and then transmit the converted DC12V or DC24V to the product. Traffic radar and video all-in-one are generally installed and used in municipal roads, where there are many electrical equipment, including traffic monitoring, traffic lights, lighting, landscape lights, etc. At the same time, the transmission lines in municipal roads are long, which can easily cause large voltage interference fluctuations at some installation points. When the peak voltage of the interference fluctuation is close to the maximum tolerance of the traffic radar and video all-in-one, it will cause the traffic radar and video all-in-one to work unstably, and in severe cases, it will cause problems such as freezing and damage.
[0005] In the prior art, the voltage interference fluctuation at the installation point is stabilized by adding a multi-stage filtering and voltage stabilizing module. However, this method requires the additional installation of a multi-stage filtering and voltage stabilizing module, which increases the cost and is not conducive to large-scale installation and deployment. Summary of the invention
[0006] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a traffic radar and video all-in-one adaptive tuning system and method, which is used to reduce voltage fluctuations and improve the working stability of the traffic radar and video all-in-one without increasing additional costs.
[0007] To achieve the above object, the present invention provides the following technical solution: a traffic radar and video integrated machine adaptive tuning system, comprising:
[0008] A power module, used for providing an input voltage, wherein the input voltage is associated with a number of corresponding standard ripple data;
[0009] A voltage conversion module, connected to the power module, for converting the input voltage into a plurality of branch voltages;
[0010] An acquisition module, connected to the voltage conversion module, for real-time acquisition of ripple data of each branch voltage, several frequency data of the traffic radar and the video all-in-one machine, and peripheral switch data;
[0011] An equation building module, connected to the voltage conversion module and the acquisition module, for building an initial state equation according to each of the branch voltages, each of the ripple data, the peripheral switch data, and each of the frequency data, wherein the initial state equation is used to output a comprehensive fluctuation value, and the comprehensive fluctuation value is used to represent the comprehensive voltage fluctuation degree of the traffic radar and the video all-in-one machine;
[0012] An adaptive tuning module is connected to the equation building module and the power supply module, and is used to obtain a plurality of ripple deviation values according to each ripple data and each standard ripple data, and then dynamically adjust each frequency data and the peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value.
[0013] Furthermore, the branch voltage includes a first system power supply voltage, a second system power supply voltage and a peripheral power supply voltage, the ripple data includes a first system ripple noise, a second system ripple noise and a peripheral ripple noise, the standard ripple data includes a first standard ripple noise, a second standard ripple noise and a peripheral standard ripple noise, and the frequency data includes CPU frequency data, DDR frequency data, NPU frequency data, and GPU frequency data.
[0014] Furthermore, the input voltage has a value range of , the value range of the first system ripple noise is , the value range of the second system ripple noise is , the value range of the peripheral ripple noise is The value range of the CPU frequency data is , the value range of the DDR frequency data is , the value range of the NPU frequency data is , the value range of GPU frequency data is , the value range of the peripheral switch data is .
[0015] Furthermore, the initial state equation is configured as: in, Used to represent the comprehensive fluctuation value, is used to indicate the first system supply voltage, Used to represent the first system ripple noise, is used to indicate the second system supply voltage, For the second system ripple noise, Used to indicate the peripheral supply voltage, Used to represent the peripheral ripple noise, Used to indicate CPU fixed-point frequency voltage data. Used to indicate DDR fixed-point frequency voltage data. Used to represent NPU fixed-point frequency voltage data, Used to represent GPU fixed-point frequency voltage data. Used to represent the peripheral switch data;
[0016] The CPU fixed-point frequency voltage data and the CPU frequency data are pre-configured with a CPU comparison relationship, and the CPU frequency data is matched according to the CPU comparison relationship to obtain the corresponding CPU fixed-point frequency voltage data;
[0017] The DDR fixed-point frequency voltage data and the DDR frequency data are pre-configured with a DDR comparison relationship, and the DDR frequency data is matched according to the DDR comparison relationship to obtain the corresponding DDR fixed-point frequency voltage data;
[0018] The NPU fixed-point frequency voltage data and the NPU frequency data are pre-configured with an NPU comparison relationship, and the NPU frequency data are matched according to the NPU comparison relationship to obtain the corresponding NPU fixed-point frequency voltage data;
[0019] The GPU fixed-point frequency voltage data and the GPU frequency data are pre-configured with a GPU comparison relationship, and the GPU frequency data is matched according to the GPU comparison relationship to obtain the corresponding GPU fixed-point frequency voltage data.
[0020] Further, the ripple deviation value includes a first ripple deviation value, a second ripple deviation value and a peripheral ripple deviation value, and the adaptive tuning module includes:
[0021] a first calculation unit, configured to obtain the first ripple deviation value by subtracting the first system ripple noise from the first standard ripple noise, obtain the second ripple deviation value by subtracting the second system ripple noise from the second standard ripple noise, and obtain the peripheral ripple deviation value by subtracting the peripheral ripple noise from the peripheral standard ripple noise;
[0022] A storage unit, used for storing associations between a plurality of first ripple deviation values and a plurality of DDR adjustment instructions, associations between a plurality of second ripple deviation values and a plurality of CPU adjustment instructions, a plurality of GPU adjustment instructions and a plurality of NPU adjustment instructions, and associations between a plurality of peripheral ripple deviation values and a plurality of peripheral adjustment instructions;
[0023] a matching unit connected to the first calculation unit and the storage unit, configured to obtain the corresponding DDR adjustment instruction according to the first ripple deviation value, to obtain the corresponding CPU adjustment instruction, the GPU adjustment instruction and the NPU adjustment instruction according to the second ripple deviation value, and to obtain the corresponding peripheral adjustment instruction according to the peripheral ripple deviation value;
[0024] an adjusting unit, connected to the matching unit, for adjusting the DDR frequency data according to the DDR adjusting instruction, adjusting the CPU frequency data, the GPU frequency data and the NPU frequency data according to the CPU adjusting instruction, the GPU adjusting instruction and the NPU adjusting instruction respectively, and adjusting the peripheral switch data according to the peripheral adjusting instruction.
[0025] Furthermore, an optimization module is included, the optimization module is connected to the adaptive tuning module, and the optimization module includes:
[0026] A detection unit, used for obtaining the first system ripple noise, the second system ripple noise and the peripheral device ripple noise at the start time and the end time of a plurality of detection cycles;
[0027] a second calculation unit, connected to the detection unit, for respectively subtracting the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the start of a cycle from the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the end of a cycle to obtain a first cycle ripple noise, a second cycle ripple noise, and a third cycle ripple noise;
[0028] an optimization unit, connected to the second calculation unit, and used to establish an optimization equation according to the first period ripple noise, the second period ripple noise and the third period ripple noise, wherein the optimization equation is used to calculate and obtain a preferred fluctuation value;
[0029] The adaptive tuning module is further used to compare the preferred fluctuation value, and when the preferred fluctuation value is greater than the comprehensive fluctuation value, dynamically adjust the frequency data and the peripheral switch data according to the ripple deviation values to reduce the preferred fluctuation value.
[0030] Furthermore, the optimization equation is configured as: Among them, is used to represent the preferred fluctuation value, 1, 2, ... n are used to represent the sequence number of the detection cycle, and there are n groups of detection cycles in total. Used to represent the first cycle ripple noise in the nth detection cycle, Used to represent the second cycle ripple noise within the nth detection cycle, Used to indicate the third cycle ripple noise within the nth detection cycle.
[0031] Furthermore, the adaptive tuning module also includes:
[0032] A recording unit, configured to record in real time the first system ripple noise, the second system ripple noise, the peripheral ripple noise, the CPU frequency data, the DDR frequency data, the NPU frequency data, the GPU frequency data, and the peripheral switch data;
[0033] An acquisition unit, used to acquire task processing volume, chip temperature data, and overall power consumption data in real time;
[0034] a disturbance calculation unit, connected to the recording unit and the acquisition unit, for introducing a disturbance correction model that has been pre-trained, and inputting the task processing amount, the chip temperature data, the overall power consumption data, the CPU frequency data, the DDR frequency data, the NPU frequency data, the GPU frequency data and the peripheral switch data into the disturbance correction model to obtain a CPU real-time disturbance value, a DDR real-time disturbance value, an NPU real-time disturbance value, a GPU real-time disturbance value and a peripheral real-time disturbance value;
[0035] a correction unit, connected to the disturbance calculation unit and the matching unit, and used for respectively correcting the CPU adjustment instruction, the GPU adjustment instruction, the NPU adjustment instruction and the peripheral adjustment instruction according to the CPU real-time disturbance value, the DDR real-time disturbance value, the NPU real-time disturbance value, the GPU real-time disturbance value and the peripheral real-time disturbance value to obtain the corrected CPU adjustment instruction, the GPU adjustment instruction, the NPU adjustment instruction and the peripheral adjustment instruction.
[0036] A method for adaptively tuning a traffic radar and video integrated machine is applied to the adaptively tuning system of the traffic radar and video integrated machine, comprising:
[0037] Step S1, the power supply module provides an input voltage, the input voltage is associated with a number of corresponding standard ripple data, and the voltage conversion module converts the input voltage into a number of branch voltages;
[0038] Step S2, the acquisition module collects the ripple data of each branch voltage, several frequency data of the traffic radar and the video all-in-one machine, and peripheral switch data in real time;
[0039] Step S3, the equation building module builds an initial state equation according to each of the branch voltages, each of the ripple data, the peripheral switch data, and each of the frequency data, wherein the initial state equation is used to output a comprehensive fluctuation value, and the comprehensive fluctuation value is used to represent the comprehensive voltage fluctuation degree of the traffic radar and the video all-in-one machine;
[0040] Step S4, the adaptive tuning module obtains a plurality of ripple deviation values according to each ripple data and each standard ripple data, and then dynamically adjusts each frequency data and the peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value.
[0041] Beneficial effects of the present invention:
[0042] The present invention first converts the input voltage of the power supply input into several branch voltages, and then collects the ripple data of each branch voltage, several frequency data of the traffic radar and video all-in-one machine, and the peripheral switch data in real time, and then establishes the initial state equation according to each ripple data, each frequency data, and the peripheral switch data to calculate the comprehensive fluctuation value used to represent the comprehensive fluctuation degree of the voltage of the traffic radar and the video all-in-one machine, so as to obtain several ripple deviation values according to each ripple data and each standard ripple data processing, and then dynamically adjusts each frequency data and the peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value. Therefore, the present invention realizes the comprehensive fluctuation of the actual voltage by simply relying on the dynamic adjustment of the frequency data and the peripheral switch data of the traffic radar and the video all-in-one machine, improves the working stability of the traffic radar and the video all-in-one machine, does not need to add additional hardware equipment, reduces the voltage stabilization cost, and is conducive to promotion. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] The present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0044] Figure 1 It is a structural schematic diagram of the traffic radar and video integrated machine adaptive tuning system in the present invention;
[0045] Figure 2 It is a schematic diagram of the relationship between load current and ripple in the present invention;
[0046] Figure 3 It is a structural diagram of the adaptive tuning module in the present invention;
[0047] Figure 4 It is a structural schematic diagram of the optimization module in the present invention;
[0048] Figure 5 It is a flow chart of the steps of the adaptive tuning method of the traffic radar and video integrated machine in the present invention.
[0049] Figure numerals: 1. power supply module; 2. voltage conversion module; 3. acquisition module; 4. equation building module; 5. adaptive tuning module; 51. first calculation unit; 52. storage unit; 53. matching unit; 54. adjustment unit; 55. acquisition unit; 56. disturbance calculation unit; 57. correction unit; 58. recording unit; 6. optimization module; 61. detection unit; 62. second calculation unit; 63. optimization unit. DETAILED DESCRIPTION
[0050] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The same parts are represented by the same reference numerals. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the accompanying drawings, and the words "bottom surface" and "top surface", "inner" and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.
[0051] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a traffic radar and video integrated machine adaptive tuning system, which can reduce voltage fluctuations and improve the working stability of the traffic radar and video integrated machine without adding additional costs, including:
[0052] The power module 1 is used to provide an input voltage, and the input voltage is associated with a number of corresponding standard ripple data;
[0053] The voltage conversion module 2 is connected to the power module 1 and is used to convert the input voltage into a plurality of branch voltages;
[0054] The acquisition module 3 is connected to the voltage conversion module 2 and is used to collect the ripple data of each branch voltage, several frequency data of the traffic radar and the video all-in-one machine, and the peripheral switch data in real time;
[0055] The equation building module 4 is connected to the voltage conversion module 2 and the acquisition module 3, and is used to establish an initial state equation according to each branch voltage, each ripple data, peripheral switch data and each frequency data. The initial state equation is used to output a comprehensive fluctuation value, and the comprehensive fluctuation value is used to represent the comprehensive voltage fluctuation degree of the traffic radar and the video all-in-one machine;
[0056] The adaptive tuning module 5 is connected to the equation building module 4 and the power supply module 1, and is used to obtain a number of ripple deviation values according to each ripple data and each standard ripple data, and then dynamically adjust each frequency data and peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value.
[0057] Specifically, in this embodiment, the power supply module 1 can be a 9VDC-32VDC programmable power supply with an external communication interface, and the voltage conversion module 2 is a DC-DC conversion circuit for converting the input voltage into 3 sets of branch voltages. Among them, the acquisition module 3 includes an oscilloscope and a data acquisition program built based on python, and the data acquisition program is pre-configured on the traffic radar and video all-in-one machine. The bandwidth of the oscilloscope is greater than 20M and has an external interface. The equation establishment module 4 can be a local calculation program built based on python, and the adaptive tuning module 5 can be a remote login adjustment program configured on a remote intelligent terminal, which is used to realize dynamic adjustment of each frequency data and peripheral switch data.
[0058] Preferably, the branch voltage includes a first system power supply voltage, a second system power supply voltage and a peripheral power supply voltage, the ripple data includes a first system ripple noise, a second system ripple noise and a peripheral ripple noise, the standard ripple data includes a first standard ripple noise, a second standard ripple noise and a peripheral standard ripple noise, and the frequency data includes CPU frequency data, DDR frequency data, NPU frequency data and GPU frequency data.
[0059] Specifically, in this embodiment, the first system power supply voltage can be a system 4V power supply voltage, which is mainly used to power the core power module 1PMIC, and is converted into 1V8 through DCDC to power peripherals such as HDMI. The second system power supply voltage can be a system 3V3 power supply voltage, which is mainly used to power the core module, and the peripheral power supply voltage can be a peripheral 5V0 power supply voltage, which is mainly used to power external devices, including serial ports, USB, etc. The first system ripple noise can be the ripple noise generated by the system 4V power supply voltage, the second system ripple noise can be the ripple noise generated by the system 3V3 power supply voltage, and the peripheral standard ripple noise can be the ripple noise generated by the peripheral 5V0 power supply voltage. The first standard ripple noise, the second standard ripple noise, and the peripheral standard ripple noise are pre-configured ripple noise standard values of the system 4V power supply voltage, the system 3V3 power supply voltage, and the peripheral 5V0 power supply voltage.
[0060] Preferably, the input voltage range is , the value range of the first system ripple noise is , the value range of the second system ripple noise is , the value range of peripheral ripple noise is , the value range of CPU frequency data is , the value range of DDR frequency data is , the value range of NPU frequency data is , the value range of GPU frequency data is , the value range of the peripheral switch data is .
[0061] Specifically, in this embodiment, the value range of NPU frequency data is , the value range of GPU frequency data is . It can be seen that the CPU frequency data has 11 adjustable gears, the DDR frequency data has 4 adjustable gears, the NPU frequency data has 9 adjustable gears, and the GPU frequency data has 9 adjustable gears.
[0062] Preferably, the initial state equation is configured as: in, Used to indicate the comprehensive fluctuation value. Used to indicate the first system supply voltage, Used to represent the first system ripple noise, Used to indicate the second system supply voltage, Used to represent the second system ripple noise, Used to indicate the peripheral power supply voltage. Used to represent peripheral ripple noise, Used to indicate CPU fixed-point frequency voltage data. Used to indicate DDR fixed-point frequency voltage data. Used to represent NPU fixed-point frequency voltage data, Used to represent GPU fixed-point frequency voltage data. Used to represent peripheral switch data;
[0063] The CPU fixed-point frequency voltage data and the CPU frequency data are pre-configured with a CPU comparison relationship, and the CPU frequency data is matched according to the CPU comparison relationship to obtain the corresponding CPU fixed-point frequency voltage data;
[0064] The DDR fixed-point frequency voltage data and the DDR frequency data are pre-configured with a DDR comparison relationship, and the DDR frequency data is matched according to the DDR comparison relationship to obtain corresponding DDR fixed-point frequency voltage data;
[0065] The NPU fixed-point frequency voltage data and the NPU frequency data are pre-configured with an NPU comparison relationship, and the NPU frequency data are matched according to the NPU comparison relationship to obtain the corresponding NPU fixed-point frequency voltage data;
[0066] The GPU fixed-point frequency voltage data and the GPU frequency data are pre-configured with a GPU comparison relationship, and the GPU frequency data is matched according to the GPU comparison relationship to obtain the corresponding GPU fixed-point frequency voltage data.
[0067] Specifically, in this embodiment, the unit of the comprehensive fluctuation value is millivolt.
[0068] Frequency Hz Voltage uV 408000000 675000 600000000 675000 816000000 675000 1008000000 675000 1200000000 675000 1416000000 675000 1608000000 700000 1800000000 762500 2016000000 837500 2208000000 912500 2400000000 962500
[0069] Frequency Hz Voltage uV 528000000 675000 1068000000 725000 1560000000 800000 2112000000 850000
[0070] Frequency Hz Voltage uV 200000000 675000 300000000 675000 400000000 675000 500000000 675000 600000000 675000 700000000 675000 800000000 700000 900000000 750000 1000000000 800000
[0071] Frequency Hz Voltage uV 200000000 675000 300000000 675000 400000000 675000 500000000 675000 600000000 675000 700000000 675000 800000000 700000 900000000 750000 1000000000 800000
[0072] In this embodiment, when the load is constant, the smaller the load current and the smaller the voltage, the smaller the ripple noise generated. Conversely, the larger the ripple noise. When the power supply voltage fluctuates, the first system ripple noise, the second system ripple noise and the peripheral ripple noise fluctuate accordingly. Assuming the load current is i, the overall operating frequency of the load is in direct proportion to the load current, such as Figure 2 As shown, when the overall operating frequency of the load decreases, the load current decreases accordingly, and the generated ripple noise also decreases accordingly. Conversely, the ripple noise increases.
[0073] When the peak-to-peak values of the first system ripple noise, the second system ripple noise and the peripheral ripple noise are close to the maximum tolerance value of the PCBA product, the product will cause unstable operation. At this time, the comprehensive fluctuation value increases. Therefore, this technical solution requires reducing the CPU frequency data, DDR frequency data, NPU frequency data, GPU frequency data and peripheral switch data while ensuring normal operation functions. It can reduce the CPU fixed-point frequency voltage data, DDR fixed-point frequency voltage data, NPU fixed-point frequency voltage data and GPU fixed-point frequency voltage data, so that the comprehensive fluctuation value of the initial state equation output is reduced accordingly.
[0074] Preferably, the ripple deviation value includes a first ripple deviation value, a second ripple deviation value and an external ripple deviation value, such as Figure 3 As shown, the adaptive tuning module 5 includes:
[0075] A first calculation unit 51 is used for subtracting the first system ripple noise from the first standard ripple noise to obtain a first ripple deviation value, subtracting the second system ripple noise from the second standard ripple noise to obtain a second ripple deviation value, and subtracting the peripheral ripple noise from the peripheral standard ripple noise to obtain a peripheral ripple deviation value;
[0076] The storage unit 52 is used to store the association relationship between the first ripple deviation values and the DDR adjustment instructions, the association relationship between the second ripple deviation values and the CPU adjustment instructions, the GPU adjustment instructions and the NPU adjustment instructions, and the association relationship between the peripheral ripple deviation values and the peripheral adjustment instructions;
[0077] A matching unit 53, connected to the first calculation unit 51 and the storage unit 52, for obtaining a corresponding DDR adjustment instruction according to the first ripple deviation value, for obtaining a corresponding CPU adjustment instruction, a GPU adjustment instruction and an NPU adjustment instruction according to the second ripple deviation value, and for obtaining a corresponding peripheral adjustment instruction according to the peripheral ripple deviation value;
[0078] The adjustment unit 54 is connected to the matching unit 53, and is used to adjust the DDR frequency data according to the DDR adjustment instruction, adjust the CPU frequency data, GPU frequency data and NPU frequency data according to the CPU adjustment instruction, GPU adjustment instruction and NPU adjustment instruction respectively, and adjust the peripheral switch data according to the peripheral adjustment instruction.
[0079] Specifically, in this embodiment, the storage unit 52 may be a non-volatile memory. The CPU adjustment instruction corresponds to one of the 11 adjustable gears of the CPU frequency data, and is used to adjust the CPU frequency data to a corresponding frequency gear; the DDR adjustment instruction corresponds to one of the 4 adjustable gears of the DDR frequency data, and is used to adjust the DDR frequency data to a corresponding frequency gear; the GPU adjustment instruction corresponds to one of the 9 adjustable gears of the GPU frequency data, and is used to adjust the GPU frequency data to a corresponding frequency gear; the NPU adjustment instruction corresponds to one of the 9 adjustable gears of the NPU frequency data, and is used to adjust the NPU frequency data to a corresponding frequency gear.
[0080] Working principle of embodiment 1:
[0081] This technical solution first converts the input voltage of the power supply input into several branch voltages, and then collects the ripple data of each branch voltage, several frequency data of the traffic radar and video all-in-one machine, and peripheral switch data in real time, and then establishes the initial state equation according to each ripple data, each frequency data, and peripheral switch data to calculate the comprehensive fluctuation value used to represent the comprehensive fluctuation degree of the voltage of the traffic radar and video all-in-one machine, so as to obtain several ripple deviation values according to the processing of each ripple data and each standard ripple data, and then dynamically adjusts each frequency data and peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value. Therefore, this technical solution realizes the comprehensive fluctuation of the actual voltage by simply relying on the dynamic adjustment of the frequency data and peripheral switch data of the traffic radar and video all-in-one machine, improves the working stability of the traffic radar and video all-in-one machine, does not require additional hardware equipment, reduces the voltage stabilization cost, and is conducive to promotion.
[0082] Example 2, reference Figure 4 , which is the second embodiment of the present invention. Different from the previous embodiment, this embodiment provides an optimization module 6, which can calculate and obtain a preferred fluctuation value to replace the comprehensive fluctuation value, and can more accurately reflect the voltage fluctuation degree. It also includes an optimization module 6, and the optimization module 6 is connected to the adaptive tuning module 5, such as Figure 4 As shown, the optimization module 6 includes:
[0083] The detection unit 61 is used to obtain the first system ripple noise, the second system ripple noise and the peripheral ripple noise at the start time and the end time of a number of detection cycles;
[0084] The second calculation unit 62 is connected to the detection unit 61 and is used to respectively subtract the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the start of the cycle from the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the end of the cycle to obtain the first cycle ripple noise, the second cycle ripple noise, and the third cycle ripple noise;
[0085] The optimization unit 63 is connected to the second calculation unit 62 and is used to establish an optimization equation according to the first cycle ripple noise, the second cycle ripple noise and the third cycle ripple noise, and the optimization equation is used to calculate and obtain a preferred fluctuation value;
[0086] The adaptive tuning module 5 is also used to compare the preferred fluctuation value. When the preferred fluctuation value is greater than the comprehensive fluctuation value, each frequency data and peripheral switch data is dynamically adjusted according to each ripple deviation value to reduce the preferred fluctuation value. Preferably, the optimization equation is configured as: Among them, is used to indicate the preferred fluctuation value, 1, 2, ... n are used to indicate the sequence number of the detection cycle, and there are n groups of detection cycles. Used to represent the first cycle ripple noise in the nth detection cycle, Used to represent the second cycle ripple noise within the nth detection cycle, Used to indicate the third cycle ripple noise within the nth detection cycle.
[0087] Specifically, in this embodiment, the unit of the preferred fluctuation value is also millivolts. According to the above optimization equation, the preferred fluctuation value is obtained by accumulating the first-cycle ripple noise, the second-cycle ripple noise and the third-cycle ripple noise in several detection cycles. The preferred fluctuation value obtained by accumulating multiple groups of first-cycle ripple noise, second-cycle ripple noise and third-cycle ripple noise can better reflect the change of ripple noise, and thus more accurately reflect the degree of voltage fluctuation.
[0088] Working principle of embodiment 2:
[0089] The optimization module is an optimization and adjustment program configured on a remote intelligent terminal. The number of detection cycles is set to three. The first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the start of the cycle are respectively subtracted from the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the end of the cycle to obtain the first cycle ripple noise, the second cycle ripple noise, and the third cycle ripple noise. When the first cycle ripple noise, the second cycle ripple noise, and the third cycle ripple noise are greater than 0, it indicates that the ripple noise value of the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the end of the cycle in the detection cycle is greater than the ripple noise value at the start of the cycle, and the voltage fluctuation becomes larger, which is not conducive to the working stability of the traffic radar and the video all-in-one machine. Therefore, it is necessary to reduce the frequency data and peripheral switch data to reduce the preferred fluctuation value. The preferred fluctuation value can better reflect the ripple fluctuation amount in several consecutive cycles than the comprehensive fluctuation value. Therefore, when the preferred fluctuation value is larger than the comprehensive fluctuation value, the preferred fluctuation value is used to replace the comprehensive fluctuation value as the judgment value for adjusting the frequency data and peripheral switch data, and then the frequency data and peripheral switch data are dynamically adjusted to reduce the preferred fluctuation value and reduce the comprehensive voltage fluctuation.
[0090] Embodiment 3 is the third embodiment of the present invention. Different from the previous embodiment, this embodiment provides a recording unit 58, an acquisition unit 55, a disturbance calculation unit 56 and a correction unit 57, which can realize the correction of CPU adjustment instructions, GPU adjustment instructions, NPU adjustment instructions and peripheral adjustment instructions, taking into account the operating temperature, task processing volume and operating power consumption while ensuring ripple stability, and improving the overall operating efficiency of the internal chip of the traffic radar and video all-in-one machine. The adaptive tuning module 5 also includes:
[0091] The recording unit 58 is used to record the first system ripple noise, the second system ripple noise, the peripheral ripple noise, the CPU frequency data, the DDR frequency data, the NPU frequency data, the GPU frequency data and the peripheral switch data in real time;
[0092] An acquisition unit 55 is used to acquire task processing amount, chip temperature data and overall power consumption data in real time;
[0093] The disturbance calculation unit 56 is connected to the recording unit 58 and the acquisition unit 55, and is used to introduce a disturbance correction model that has been pre-trained, and input the task processing amount, chip temperature data, overall power consumption data, CPU frequency data, DDR frequency data, NPU frequency data, GPU frequency data and peripheral switch data into the disturbance correction model to obtain the CPU real-time disturbance value, DDR real-time disturbance value, NPU real-time disturbance value, GPU real-time disturbance value and peripheral real-time disturbance value;
[0094] The correction unit 57 is connected to the disturbance calculation unit 56 and the matching unit 53, and is used to correct the CPU adjustment instruction, the GPU adjustment instruction, the NPU adjustment instruction and the peripheral adjustment instruction according to the CPU real-time disturbance value, the DDR real-time disturbance value, the NPU real-time disturbance value, the GPU real-time disturbance value and the peripheral real-time disturbance value, respectively, to obtain the corrected CPU adjustment instruction, GPU adjustment instruction, NPU adjustment instruction and peripheral adjustment instruction.
[0095] Working principle of embodiment 3:
[0096] The recording unit 58 records the first system ripple noise, the second system ripple noise, the peripheral ripple noise, the CPU frequency data, the DDR frequency data, the NPU frequency data, the GPU frequency data and the peripheral switch data in real time during the detection and debugging process, and generates a data log. The acquisition unit 55 obtains the task data volume of the CPU, DDR, NPU, GPU and peripheral devices through data acquisition, and simultaneously obtains the chip temperature data and the overall temperature data of the chip. The disturbance correction model is obtained based on the GBDT model training. The training process of the disturbance correction model is to take several training task processing volumes, training chip temperature data, training overall power consumption data, training CPU frequency data, training DDR frequency data, training NPU frequency data, training GPU frequency data and training peripheral switch data in the training database as input, and take the corresponding several training CPU real-time disturbance values, training DDR real-time disturbance values, training NPU real-time disturbance values, training GPU real-time disturbance values and training peripheral real-time disturbance values as output, train the original GBDT model several times, and ensure that its prediction accuracy reaches 92% before outputting it as a disturbance correction model. The CPU real-time disturbance value, DDR real-time disturbance value, NPU real-time disturbance value, GPU real-time disturbance value and peripheral real-time disturbance value predicted by the disturbance correction model comprehensively consider the impact of chip temperature, overall power consumption and task processing volume on the demand for each frequency data, and make fine adjustments based on the original CPU adjustment instructions, GPU adjustment instructions, NPU adjustment instructions and peripheral adjustment instructions, so that the final adjusted CPU frequency data, GPU frequency data, NPU frequency data, and peripheral switch data are more in line with the actual operating scenarios, taking into account the operating temperature, task processing volume and operating power consumption while ensuring ripple stability, thereby improving the overall operating efficiency of the internal chips of the traffic radar and video all-in-one machine.
[0097] A traffic radar and video integrated machine adaptive tuning method is applied to the above-mentioned traffic radar and video integrated machine adaptive tuning system, such as Figure 5 As shown, including:
[0098] Step S1, the power module 1 provides an input voltage, the input voltage is associated with a number of corresponding standard ripple data, and the voltage conversion module 2 converts the input voltage into a number of branch voltages;
[0099] Step S2, the acquisition module 3 collects the ripple data of each branch voltage, several frequency data of the traffic radar and the video all-in-one machine, and the peripheral switch data in real time;
[0100] Step S3, the equation building module 4 builds an initial state equation according to each branch voltage, each ripple data, peripheral switch data and each frequency data, and the initial state equation is used to output a comprehensive fluctuation value, and the comprehensive fluctuation value is used to represent the comprehensive voltage fluctuation degree of the traffic radar and the video all-in-one machine;
[0101] In step S4, the adaptive tuning module 5 obtains a number of ripple deviation values according to the ripple data and the standard ripple data, and then dynamically adjusts the frequency data and the peripheral switch data according to the ripple deviation values to reduce the comprehensive fluctuation value.
[0102] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.
Claims
1. A traffic radar and video integrated machine adaptive tuning system, characterized in that: include: A power supply module (1) is used to provide an input voltage, wherein the input voltage is associated with a corresponding number of standard ripple data; A voltage conversion module (2), connected to the power module (1), and used for converting the input voltage into a plurality of branch voltages; A collection module (3) connected to the voltage conversion module (2) for collecting ripple data of each branch voltage, a number of frequency data of the traffic radar and the video all-in-one machine, and peripheral switch data in real time; An equation establishing module (4) is connected to the voltage conversion module (2) and the acquisition module (3), and is used to establish an initial state equation according to each branch voltage, each ripple data, the peripheral switch data, and each frequency data, wherein the initial state equation is used to output a comprehensive fluctuation value, and the comprehensive fluctuation value is used to represent the comprehensive voltage fluctuation degree of the traffic radar and the video integrated machine; An adaptive tuning module (5) is connected to the equation building module (4) and the power supply module (1) and is used to obtain a plurality of ripple deviation values according to the processing of each ripple data and each standard ripple data, and then dynamically adjust each frequency data and the peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value.
2. The traffic radar and video integrated machine adaptive tuning system according to claim 1 is characterized in that: The branch voltage includes a first system power supply voltage, a second system power supply voltage and a peripheral power supply voltage, the ripple data includes a first system ripple noise, a second system ripple noise and a peripheral ripple noise, the standard ripple data includes a first standard ripple noise, a second standard ripple noise and a peripheral standard ripple noise, and the frequency data includes CPU frequency data, DDR frequency data, NPU frequency data, and GPU frequency data.
3. The traffic radar and video integrated machine adaptive tuning system according to claim 2 is characterized in that: The input voltage range is , the value range of the first system ripple noise is , the value range of the second system ripple noise is , the value range of the peripheral ripple noise is The value range of the CPU frequency data is , the value range of the DDR frequency data is , the value range of the NPU frequency data is , the value range of the GPU frequency data is , the value range of the peripheral switch data is .
4. The traffic radar and video integrated machine adaptive tuning system according to claim 3 is characterized in that: The initial state equation is configured as: in, Used to represent the comprehensive fluctuation value, is used to indicate the first system supply voltage, Used to represent the first system ripple noise, is used to indicate the second system supply voltage, For the second system ripple noise, Used to indicate the peripheral supply voltage, Used to represent the peripheral ripple noise, Used to indicate CPU fixed-point frequency voltage data. Used to indicate DDR fixed-point frequency voltage data. Used to represent NPU fixed-point frequency voltage data, Used to represent GPU fixed-point frequency voltage data. Used to represent the peripheral switch data; The CPU fixed-point frequency voltage data and the CPU frequency data are pre-configured with a CPU comparison relationship, and the CPU frequency data is matched according to the CPU comparison relationship to obtain the corresponding CPU fixed-point frequency voltage data; The DDR fixed-point frequency voltage data and the DDR frequency data are pre-configured with a DDR comparison relationship, and the DDR frequency data is matched according to the DDR comparison relationship to obtain the corresponding DDR fixed-point frequency voltage data; The NPU fixed-point frequency voltage data and the NPU frequency data are pre-configured with an NPU comparison relationship, and the NPU frequency data are matched according to the NPU comparison relationship to obtain the corresponding NPU fixed-point frequency voltage data; The GPU fixed-point frequency voltage data and the GPU frequency data are pre-configured with a GPU comparison relationship, and the GPU frequency data is matched according to the GPU comparison relationship to obtain the corresponding GPU fixed-point frequency voltage data.
5. The traffic radar and video integrated machine adaptive tuning system according to claim 2 is characterized in that: The ripple deviation value includes a first ripple deviation value, a second ripple deviation value and a peripheral ripple deviation value, and the adaptive tuning module (5) includes: A first calculation unit (51) is used for subtracting the first system ripple noise from the first standard ripple noise to obtain the first ripple deviation value, subtracting the second system ripple noise from the second standard ripple noise to obtain the second ripple deviation value, and subtracting the peripheral ripple noise from the peripheral standard ripple noise to obtain the peripheral ripple deviation value; A storage unit (52) is used to store associations between a plurality of first ripple deviation values and a plurality of DDR adjustment instructions, associations between a plurality of second ripple deviation values and a plurality of CPU adjustment instructions, a plurality of GPU adjustment instructions and a plurality of NPU adjustment instructions, and associations between a plurality of peripheral ripple deviation values and a plurality of peripheral adjustment instructions; a matching unit (53), connected to the first calculation unit (51) and the storage unit (52), and configured to obtain the corresponding DDR adjustment instruction according to the first ripple deviation value, to obtain the corresponding CPU adjustment instruction, the GPU adjustment instruction and the NPU adjustment instruction according to the second ripple deviation value, and to obtain the corresponding peripheral adjustment instruction according to the peripheral ripple deviation value; An adjustment unit (54) is connected to the matching unit (53) and is used to adjust the DDR frequency data according to the DDR adjustment instruction, adjust the CPU frequency data, the GPU frequency data and the NPU frequency data according to the CPU adjustment instruction, the GPU adjustment instruction and the NPU adjustment instruction respectively, and adjust the peripheral switch data according to the peripheral adjustment instruction.
6. The traffic radar and video integrated machine adaptive tuning system according to claim 4, characterized in that: The system further comprises an optimization module (6), wherein the optimization module (6) is connected to the adaptive tuning module (5), and the optimization module (6) comprises: A detection unit (61) is used to obtain the first system ripple noise, the second system ripple noise and the peripheral device ripple noise at the start time and the end time of a plurality of detection cycles; a second calculation unit (62), connected to the detection unit (61), and used to respectively perform a subtraction of the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the start of a cycle with the first system ripple noise, the second system ripple noise, and the peripheral ripple noise at the end of a cycle to obtain a first cycle ripple noise, a second cycle ripple noise, and a third cycle ripple noise; an optimization unit (63), connected to the second calculation unit (62), and used to establish an optimization equation according to the first period ripple noise, the second period ripple noise and the third period ripple noise, wherein the optimization equation is used to calculate and obtain a preferred fluctuation value; The adaptive tuning module (5) is further used to compare the preferred fluctuation value, and when the preferred fluctuation value is greater than the comprehensive fluctuation value, dynamically adjust the frequency data and the peripheral switch data according to the ripple deviation values to reduce the preferred fluctuation value.
7. The traffic radar and video integrated machine adaptive tuning system according to claim 6, characterized in that: The optimization equation is configured as: Among them, is used to represent the preferred fluctuation value, 1, 2, ... n are used to represent the sequence number of the detection cycle, and there are n groups of detection cycles in total. Used to represent the first cycle ripple noise in the nth detection cycle, Used to represent the second cycle ripple noise within the nth detection cycle, Used to indicate the third cycle ripple noise within the nth detection cycle.
8. The traffic radar and video integrated machine adaptive tuning system according to claim 5, characterized in that: The adaptive tuning module (5) further comprises: A recording unit (58), used for recording in real time the first system ripple noise, the second system ripple noise, the peripheral ripple noise, the CPU frequency data, the DDR frequency data, the NPU frequency data, the GPU frequency data and the peripheral switch data; An acquisition unit (55), used for acquiring task processing amount, chip temperature data and overall power consumption data in real time; A disturbance calculation unit (56) is connected to the recording unit (58) and the acquisition unit (55), and is used to introduce a disturbance correction model that has been trained in advance, and input the task processing amount, the chip temperature data, the overall power consumption data, the CPU frequency data, the DDR frequency data, the NPU frequency data, the GPU frequency data and the peripheral switch data into the disturbance correction model to obtain a CPU real-time disturbance value, a DDR real-time disturbance value, an NPU real-time disturbance value, a GPU real-time disturbance value and a peripheral real-time disturbance value; A correction unit (57) is connected to the disturbance calculation unit (56) and the matching unit (53), and is used to correct the CPU adjustment instruction, the GPU adjustment instruction, the NPU adjustment instruction and the peripheral adjustment instruction according to the CPU real-time disturbance value, the DDR real-time disturbance value, the NPU real-time disturbance value, the GPU real-time disturbance value and the peripheral real-time disturbance value, respectively, to obtain the corrected CPU adjustment instruction, the GPU adjustment instruction, the NPU adjustment instruction and the peripheral adjustment instruction.
9. A method for adaptively tuning a traffic radar and video integrated device, applied to the adaptively tuning system for a traffic radar and video integrated device as claimed in any one of claims 1 to 8, characterized in that: include: Step S1, the power supply module (1) provides an input voltage, the input voltage is associated with a number of corresponding standard ripple data, and the voltage conversion module (2) converts the input voltage into a number of branch voltages; Step S2, the acquisition module (3) collects the ripple data of each branch voltage, a number of frequency data of the traffic radar and the video all-in-one machine, and peripheral switch data in real time; Step S3, the equation building module (4) builds an initial state equation according to each of the branch voltages, each of the ripple data, the peripheral switch data, and each of the frequency data, wherein the initial state equation is used to output a comprehensive fluctuation value, and the comprehensive fluctuation value is used to represent the comprehensive voltage fluctuation degree of the traffic radar and the video integrated device; Step S4, the adaptive tuning module (5) obtains a plurality of ripple deviation values according to each ripple data and each standard ripple data, and then dynamically adjusts each frequency data and the peripheral switch data according to each ripple deviation value to reduce the comprehensive fluctuation value.
Citation Information
Patent Citations
System and method for mitigating radar interference using clusters
CN115639527A
Mobile vehicle positioning method and system
CN117029840A
