A logistics vehicle dynamic scheduling method and system thereof

Abstract

The application discloses a kind of logistics vehicle dynamic scheduling method and system, it is related to logistics transport technical field, including the following steps: S1, real-time multi-source sensing data;S2, calculate vehicle power attenuation coefficient and driver physiological load index, construct three-dimensional coupling attenuation model;S3, according to the model, each distribution node is carried out accessibility evaluation, when oxygen saturation is lower than threshold value or power attenuation coefficient is higher than threshold value, trigger scheduling reconstruction, and simultaneously adjust vehicle-mounted cold chain equipment temperature control parameter to compensate in-transit time variation;S4, continuously update model and iterative optimization scheduling decision during driving process until task is completed.The application constructs three-dimensional coupling attenuation model, links highland environment parameter, driver real-time physiological state and vehicle power performance with linkage quantization, and initiatively triggers scheduling reconstruction based on oxygen saturation threshold value and power attenuation threshold value, realizes the collaborative optimization of scheduling decision and cold chain temperature control under highland complex environment.

Description

Technical Field

[0001] This invention relates to the field of logistics transportation technology, and in particular to a method and system for dynamic scheduling of logistics vehicles. Background Technology

[0002] As cold chain logistics extends to remote areas, the demand for transporting fresh livestock products from high-altitude pastoral areas is increasing. However, high-altitude pastoral areas are characterized by high altitude, low air pressure, complex road conditions, and weak communication coverage, posing unique challenges to the dynamic scheduling of logistics vehicles.

[0003] Existing dynamic scheduling methods for logistics vehicles primarily rely on traffic conditions, cargo demand, vehicle location, and general road conditions for route planning and real-time adjustments. These scheduling models typically assume stable vehicle power performance and constant driver physiological state. However, in high-altitude pastoral environments, atmospheric pressure and oxygen levels decrease significantly with increasing altitude, leading to reduced engine intake efficiency and power output. Simultaneously, drivers experience decreased blood oxygen saturation, abnormally high heart rates, and a sharp increase in physiological load, limiting their ability to operate continuously. These environmental, physiological, and power factors exhibit a significant synergistic attenuation effect: the higher the altitude, the weaker the vehicle's power and the worse the driver's physiological state, both of which jointly constrain the vehicle's actual reach and operating time.

[0004] However, existing scheduling methods do not couple and model high-altitude environmental parameters, drivers' real-time physiological states, and vehicle power performance. This makes it impossible to quantitatively assess the linkage effect of environmental changes on actual vehicle capacity and drivers' continuous working ability, resulting in a serious disconnect between scheduling plans and on-site execution capabilities. When drivers experience altitude sickness or vehicles lack power, existing systems can only provide post-event alarms and cannot incorporate physiological safety thresholds and power attenuation thresholds into scheduling decisions in advance, nor can they proactively trigger route reconstruction, rest node insertion, and time window correction. Furthermore, existing cold chain scheduling and vehicle scheduling are independent of each other. When transit time changes due to physiological or power factors, cold chain temperature control parameters cannot be adjusted in a coordinated manner, which can easily lead to cold chain disruptions and a decline in cargo quality.

[0005] Therefore, there is an urgent need for a dynamic vehicle scheduling method that can couple the attenuation relationship between the plateau environment, driver physiology, and vehicle power, and achieve coordinated optimization of scheduling decisions and cold chain temperature control. Summary of the Invention

[0006] The purpose of this invention is to provide a dynamic scheduling method and system for logistics vehicles to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamic scheduling of logistics vehicles, comprising the following steps:

[0008] S1. The vehicle uses onboard environmental sensors to obtain the current road section's altitude, atmospheric pressure, and ambient oxygen content in real time; wearable devices can be used to obtain the driver's blood oxygen saturation and heart rate in real time; and the vehicle's CAN bus can be used to obtain the engine's output power and output torque in real time.

[0009] S2. Based on the altitude, atmospheric pressure and ambient oxygen content, calculate the vehicle power attenuation coefficient and the driver's physiological load index respectively. Based on the vehicle power attenuation coefficient and the driver's physiological load index, construct a three-dimensional coupled attenuation model to characterize the linkage attenuation relationship between the vehicle's actual carrying capacity and the driver's continuous working ability.

[0010] S3. Based on the three-dimensional coupled attenuation model, the accessibility of each delivery node on the preset delivery route is evaluated. When the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold, the scheduling reconstruction is triggered. The scheduling reconstruction includes replanning the order of delivery nodes, adjusting the upper limit of the road segment driving speed, inserting rest nodes in the road and correcting the delivery time window. According to the reconstructed scheduling scheme, the temperature control parameters of the vehicle cold chain equipment are adjusted synchronously to compensate for the change in the in-transit time.

[0011] S4. During the driving process, the multi-source sensing data is continuously collected, the three-dimensional coupling attenuation model is updated, and the scheduling decision is iteratively optimized based on the updated model until all delivery tasks are completed.

[0012] Preferably, the calculation of the vehicle power attenuation coefficient includes:

[0013] The baseline attenuation rate is obtained by querying the pre-stored dynamic attenuation mapping table based on the altitude.

[0014] The baseline attenuation rate is adjusted based on the atmospheric pressure and ambient oxygen content to obtain the vehicle power attenuation coefficient.

[0015] Preferably, the calculation of the driver's physiological workload index includes:

[0016] The blood oxygen saturation and heart rate data are normalized to obtain normalized blood oxygen values ​​and normalized heart rate values.

[0017] The blood oxygen weighting coefficient and heart rate weighting coefficient are determined based on the altitude.

[0018] The driver's physiological load index is calculated based on a weighted summation.

[0019] Preferably, the accessibility assessment includes:

[0020] Based on the current vehicle location, delivery node sequence, and maximum speed limit of the road section, the estimated arrival time of the vehicle at each delivery node is determined.

[0021] Based on the three-dimensional coupled attenuation model, the predicted values ​​of the power attenuation coefficient and the physiological load index are estimated when the vehicle reaches each delivery node.

[0022] When the predicted value exceeds the preset safety threshold, the corresponding delivery node is marked as an unreachable node, and a secondary scheduling reconfiguration is triggered.

[0023] Preferably, the insertion of the in-transit rest node includes:

[0024] Based on the changing trend of the driver's physiological load index, the estimated time and geographical location when the driver reaches the fatigue threshold are predicted.

[0025] Within a preset range near the estimated geographical location, service points with oxygen supply conditions are selected as candidate rest nodes.

[0026] The candidate rest nodes are inserted into the delivery path, and the arrival times and corresponding temperature control parameters of subsequent delivery nodes are recalculated.

[0027] Preferably, the synchronous adjustment of the temperature control parameters of the vehicle-mounted cold chain equipment includes:

[0028] Calculate the change in transit time between each delivery node based on the reconstructed scheduling scheme;

[0029] Calculate the temperature control compensation value based on the change in transit time and the cold chain preservation requirements corresponding to the product category;

[0030] The set temperature and cooling power output of the vehicle-mounted cold chain equipment are adjusted according to the temperature control compensation value.

[0031] Preferably, in step S3, the triggered scheduling reconfiguration further includes:

[0032] When the rate of change of altitude of the road section ahead is detected to exceed the preset slope threshold, the increment of the dynamic attenuation coefficient of the road section ahead is pre-calculated.

[0033] The vehicle's ability to climb hills is determined based on the increase in the power attenuation coefficient.

[0034] When the judgment result is not met, the road section ahead is marked as a no-entry section, and an alternative route is planned. At the same time, the cold chain temperature control parameters are recalculated based on the change in the length of the alternative route.

[0035] Preferably, step S4 further includes:

[0036] Record the triggering conditions for each scheduling refactoring, the path differences before and after the refactoring, and the actual execution deviation;

[0037] The reinforcement learning model is trained based on the aforementioned triggering conditions, path differences, and actual execution deviations.

[0038] The trained reinforcement learning model is used to predict the scheduling risk probability of future road segments, and pre-scheduling reconstruction is triggered in advance when the scheduling risk probability exceeds a preset threshold.

[0039] Preferably, step S1 further includes:

[0040] The satellite communication module is used to obtain real-time road conditions and weather warnings for each delivery node within a pre-defined area of ​​the pastoral region.

[0041] The real-time road traffic status and weather warning information are input as external constraints into the three-dimensional coupled attenuation model to correct the accessibility assessment results.

[0042] On the other hand, the present invention also provides a dynamic scheduling system for logistics vehicles, implementing the dynamic scheduling method for logistics vehicles described in any of the above claims, the system comprising:

[0043] The multi-source sensing layer includes onboard environmental sensors, driver wearable devices, and the vehicle CAN bus interface, used to collect data in real time, such as altitude, atmospheric pressure, ambient oxygen content, blood oxygen saturation, heart rate, engine output power, and output torque.

[0044] The model building layer is used to calculate the vehicle power attenuation coefficient and the driver's physiological load index based on the data collected by the multi-source perception layer, and to build a three-dimensional coupled attenuation model.

[0045] The scheduling decision layer is used to assess the accessibility of each delivery node on the preset delivery route based on the three-dimensional coupled attenuation model, and triggers scheduling reconstruction when the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold. It outputs the reconstructed delivery node sequence, the upper limit of the road segment driving speed, the rest nodes in the way and the delivery time window, and generates cold chain temperature control parameter adjustment instructions.

[0046] A closed-loop optimization layer is used to continuously receive data from the multi-source perception layer during driving, update the three-dimensional coupling attenuation model, and iteratively optimize scheduling decisions.

[0047] The vehicle-mounted cold chain control module is used to receive the cold chain temperature control parameter adjustment command and execute the adjustment of the set temperature and cooling power output.

[0048] The technical effects and advantages of this invention are as follows:

[0049] This invention constructs a three-dimensional coupled attenuation model of environment, physiology, and power, which links and quantifies plateau environmental parameters, real-time physiological state of drivers, and vehicle power performance. Based on blood oxygen saturation threshold and power attenuation threshold, it actively triggers scheduling reconstruction, solving the problem of the disconnect between scheduling plan and actual execution capability in plateau areas in existing technologies. It effectively avoids delivery delays and safety hazards caused by driver altitude sickness or insufficient vehicle power, and significantly improves the delivery reliability and operational safety of cold chain logistics in plateau pastoral areas.

[0050] This invention, while triggering scheduling reconfiguration, adjusts the set temperature and refrigeration power output of the vehicle-mounted cold chain equipment in conjunction with the change in transit time after reconfiguration. This achieves coordinated optimization of vehicle dynamic scheduling and cold chain temperature control parameters, effectively compensating for the adverse effects of extended transit time caused by factors such as rest stop insertion and route detours on the freshness and quality of goods, and ensuring the overall quality control of cold chain logistics in plateau pastoral areas.

[0051] This invention continuously updates the three-dimensional coupling attenuation model through a closed-loop feedback mechanism and trains a reinforcement learning model based on historical scheduling data to achieve pre-schedule reconstruction. This enables the scheduling strategy to adapt to the complex and ever-changing environmental characteristics of plateau pastoral areas, gradually reducing the frequency of scheduling reconstruction and execution deviation, and significantly improving the scheduling robustness and long-term operating efficiency of the system in scenarios with weak communication, high altitude, and uncertain road conditions. Attached Figure Description

[0052] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention, but do not constitute a limitation thereof. In the drawings:

[0053] Figure 1 This is the overall flowchart of the dynamic scheduling of logistics vehicles in this invention.

[0054] Figure 2 This is a flowchart of steep slope prediction and detour in the scheduling and reconstruction process of this invention.

[0055] Figure 3 This is a schematic diagram of the pre-scheduled reconfiguration triggering structure of the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] This invention provides, for example Figures 1-3The method for dynamic scheduling of logistics vehicles shown includes the following steps:

[0058] S1. Real-time environmental perception data such as altitude, atmospheric pressure and oxygen content of the current road section are obtained through on-board environmental sensors. Real-time physiological perception data of driver's blood oxygen saturation and heart rate are obtained through wearable devices. Real-time vehicle condition perception data of engine output power and output torque are obtained through vehicle CAN bus.

[0059] S2. Based on environmental perception data, physiological perception data and vehicle condition perception data, calculate the vehicle power attenuation coefficient and the driver's physiological load index respectively. Based on the vehicle power attenuation coefficient and the driver's physiological load index, construct a three-dimensional coupled attenuation model to characterize the linkage attenuation relationship between the vehicle's actual carrying capacity and the driver's continuous working ability.

[0060] S3. Based on the three-dimensional coupled attenuation model, the accessibility of each delivery node on the preset delivery route is assessed. When the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold, the scheduling is reconfigured. The scheduling reconfiguration includes replanning the order of delivery nodes, adjusting the upper limit of the road segment driving speed, inserting rest nodes in the road and correcting the delivery time window. According to the reconfigured scheduling plan, the temperature control parameters of the vehicle cold chain equipment are adjusted synchronously to compensate for the change in the in-transit time.

[0061] S4. During the journey, multi-source perception data is continuously collected, the three-dimensional coupled attenuation model is updated, and the scheduling decision is iteratively optimized based on the updated model until all delivery tasks are completed.

[0062] In practice, the vehicle power attenuation coefficient η is used to quantify the degree of power attenuation caused by the low-oxygen and low-pressure environment at high altitudes. Its value ranges from 0 to 1, with smaller values ​​indicating more severe power attenuation. A power attenuation mapping table is pre-stored in the vehicle's storage unit, obtained through bench testing or high-altitude calibration of the actual vehicle. The mapping table uses altitude H as an index and records the engine output power at different altitudes relative to the rated power P at sea level. o In practical applications, if the current altitude H is between two calibration points, the attenuation ratio is calculated using linear interpolation. o Based on the current atmospheric pressure P a and the baseline decay rate η of ambient oxygen content (O2). o After correction, the vehicle power attenuation coefficient η is obtained. Then, the real-time collected blood oxygen saturation (SpO2) and heart rate (HR) are normalized, and the blood oxygen weighting coefficient ω is dynamically determined based on the current altitude (H). s With heart rate weighting coefficient ω hBased on the blood oxygen weighting coefficient and heart rate weighting coefficient, the normalized oxygen saturation and heart rate are weighted and summed to obtain the driver's physiological load index L. In this embodiment, the blood oxygen weighting coefficient and heart rate weighting coefficient are specifically determined as follows: when H < 2500m: ω s = 0.4, ω h =0.6; When 2500m≤H<3500m: ω s =0.5, ω h =0.5; when 3500m≤H<4500m: ω s =0.6, ω h =0.4; when H≥4500m: ω s =0.7, ω h =0.3. A three-dimensional coupled attenuation model is constructed using the vehicle power attenuation coefficient η, the driver's physiological load index L, and altitude H as three dimensions. The model's expression is: Wherein, γ1 is the power attenuation weight, which is 0.35 in this embodiment; γ2 is the physiological load weight, which is 0.45 in this embodiment; γ3 is the altitude environment weight, which is 0.20 in this embodiment; Hmax is the maximum operating altitude set by the system, which is 6000m in this embodiment. The closer C is to 1, the weaker the vehicle's actual ability to perform delivery tasks, thus characterizing the linkage attenuation relationship between the vehicle's actual carrying capacity and the driver's continuous working ability. Based on the constructed three-dimensional coupled attenuation model, the predicted values ​​of the power attenuation coefficient and physiological load index at each delivery node are estimated to conduct accessibility assessment. When the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold, the scheduling reconfiguration is triggered, the order of delivery nodes is replanned, the upper limit of the road segment driving speed is adjusted, service points with oxygen supply conditions near the estimated geographical location are selected as rest nodes to be inserted into the delivery route, and the delivery time window is corrected. At the same time, the temperature control compensation value is calculated based on the change in the travel time after the reconfiguration and the cold chain preservation requirements corresponding to the goods category, and the set temperature and cooling power output of the on-board cold chain equipment are adjusted accordingly. In addition, multi-source perception data is continuously collected during the driving process to update the three-dimensional coupled attenuation model, and the triggering conditions of each scheduling reconfiguration, the path differences before and after the reconfiguration, and the actual execution deviation are recorded to train the reinforcement learning model to predict the scheduling risk probability of future road segments. When the risk probability exceeds the preset threshold, pre-scheduling reconfiguration is triggered in advance to iteratively optimize the scheduling decision.

[0063] By using a three-dimensional coupled attenuation model to quantify the linkage between the plateau environment, driver physiology, and vehicle power, and actively triggering scheduling reconfiguration and cold chain temperature control adjustments based on safety thresholds, the system effectively solves the problem of the disconnect between scheduling plans and actual execution capabilities on the plateau. This avoids delivery delays and safety hazards caused by driver altitude sickness or insufficient vehicle power, significantly improving the delivery reliability and operational safety of cold chain logistics in plateau pastoral areas. At the same time, through closed-loop feedback and reinforcement learning pre-scheduling, the system can adapt to the complex and ever-changing environmental characteristics of the plateau, gradually reducing the frequency of scheduling reconfiguration and execution deviations, ensuring controllable quality throughout the cold chain, and improving the system's scheduling robustness and long-term operational efficiency.

[0064] In a preferred embodiment, step S1 further includes: obtaining real-time road traffic status and weather warning information of each delivery node within a preset range of the pastoral area through a satellite communication module; and inputting the real-time road traffic status and weather warning information as external constraints into a three-dimensional coupled attenuation model to correct the accessibility assessment results.

[0065] In specific implementation, the satellite communication module uses BeiDou short message service or Iridium communication link to request road conditions around each delivery node along the current delivery route and backup routes from the remote scheduling platform at preset time intervals (e.g., every 10 minutes). This includes information on whether landslides, road icing, temporary traffic control, or pasture road blockages have occurred. Simultaneously, it acquires meteorological warnings, including warnings of extreme weather such as strong winds, blizzards, heavy fog, or rapid temperature drops within the next two hours. The scheduling platform encodes this information as constraint factors and inputs them into the environmental parameter dimension of the three-dimensional coupled attenuation model to dynamically correct the vehicle power attenuation coefficient output by the model. For example, when there is snow or slippery conditions on the road ahead, a road surface adhesion coefficient reduction factor is added to the original power attenuation coefficient; when a strong wind warning is received, an air resistance increment factor is added. The revised model recalculates the estimated arrival time of each delivery node and the corresponding predicted value of the driver's physiological load index. If the revised predicted value exceeds the safety threshold, the pre-schedule reconfiguration is triggered in advance, including detouring to alternative nodes with good road conditions, adjusting the route order, or inserting rest and waiting nodes in advance in safe areas to avoid the risk of superimposed adverse weather and poor road conditions.

[0066] By coupling external environmental constraints obtained from satellite communication into the three-dimensional attenuation model in real time, the lack of scheduling information in the ground communication blind spots of plateau pastoral areas is compensated for. This allows accessibility assessment to no longer rely solely on historical road conditions and static map data, but to integrate real-time dynamic road and meteorological risk factors. This significantly improves the foresight of scheduling decisions in response to sudden road conditions and extreme weather, reduces the risk of vehicles being forced to turn back or be stranded due to mistakenly entering restricted road sections, and avoids cold chain temperature control failures caused by emergency detours, further enhancing the safety and timeliness of cold chain logistics scheduling in plateau pastoral areas.

[0067] In a preferred embodiment, the calculation of the vehicle power attenuation coefficient includes: obtaining a baseline attenuation rate by querying a pre-stored power attenuation mapping table based on altitude; and performing environmental correction on the baseline attenuation rate based on atmospheric pressure and ambient oxygen content to obtain the vehicle power attenuation coefficient. The calculation of the driver's physiological load index includes: normalizing blood oxygen saturation and heart rate data to obtain normalized blood oxygen and heart rate values; determining blood oxygen weighting coefficients and heart rate weighting coefficients based on altitude; and calculating the driver's physiological load index based on a weighted summation.

[0068] In practice, the system incorporates a built-in power attenuation mapping table, calibrated using the percentage of rated power loss measured at different altitudes (from 0 meters to 5000 meters, in 200-meter intervals) during engine bench tests. The calculation uses the current altitude as an index and employs linear interpolation to obtain the baseline attenuation rate. Subsequently, pressure correction factors and oxygen content correction factors are constructed using the ratio of current atmospheric pressure to standard atmospheric pressure (101.325 kPa) and the ratio of ambient oxygen content to sea-level oxygen content (approximately 20.95%), respectively. These two factors are then multiplied by the baseline attenuation rate to obtain the comprehensive vehicle power attenuation coefficient. Meanwhile, blood oxygen saturation (normal value 95%-100%) and heart rate (normal value 60-100 beats / min) were transformed to the range [0, 1] using the range normalization formula, where the lower the blood oxygen saturation, the higher the normalized value, and the higher the heart rate, the higher the normalized value. The weighting coefficients were dynamically determined according to the current altitude. When the altitude was below 2500 meters, the weights of blood oxygen and heart rate were both set to 0.5. For every 500 meters increase in altitude, the weight of blood oxygen increased by 0.05, and the weight of heart rate decreased accordingly, to highlight the main physiological stress effect of the high-altitude hypoxic environment. Finally, the physiological load index was calculated using the weighted summation formula, with a value range of [0, 1]. The larger the value, the heavier the load.

[0069] By comprehensively considering three factors—altitude, air pressure, and oxygen content—the power attenuation coefficient more accurately reflects the engine's actual power output capacity than the single altitude lookup table method. The physiological load index adopts a dynamic weighting strategy, enabling the evaluation model to reasonably reflect the dominant changes in blood oxygen and heart rate across different altitude ranges. This provides a more accurate decision-making basis for scheduling reconfiguration that is closer to the actual human tolerance state, effectively reducing the misjudgment rate.

[0070] In a preferred embodiment, the accessibility assessment includes: estimating the arrival time of the vehicle at each delivery node based on the current vehicle location, delivery node sequence, and road segment speed limit; estimating the predicted values ​​of the power attenuation coefficient and physiological load index when arriving at each delivery node based on a three-dimensional coupled attenuation model; and marking the corresponding delivery node as an unreachable node and triggering a secondary scheduling reconfiguration when the predicted values ​​exceed a preset safety threshold.

[0071] In practice, the dispatch platform uses the current GPS coordinates of the vehicle as the starting point and extracts the distance between each delivery node and the preset speed limit for that segment along the pre-defined delivery route (this speed limit has been initially corrected based on altitude segmentation). Using a shortest path algorithm combined with real-time traffic conditions, the estimated arrival time from the current location to the next delivery node is calculated sequentially, and then extrapolated to subsequent nodes. Simultaneously, the current altitude, atmospheric pressure, ambient oxygen content, driver's blood oxygen saturation, heart rate, engine output power, and output torque are used as initial state variables for the three-dimensional coupled attenuation model. Combined with the altitude change trend along the path (such as the gradient and altitude sequence of subsequent segments), the predicted power attenuation coefficient and driver physiological load index are extrapolated to estimate the vehicle's arrival time at each delivery node. The preset safety thresholds include an upper limit for the power attenuation coefficient (e.g., 0.65, indicating that when the actual output power of the engine is less than 65% of the rated power, it is considered unacceptable) and an upper limit for the physiological load index (e.g., 0.75, indicating that the driver is close to the critical point of moderate altitude sickness). The two thresholds can be set independently, or the union of the two exceeding the limits can be used as the unreachability criterion. When any predicted value exceeds the corresponding threshold, the delivery node is marked as an unreachable node, and a secondary scheduling reconfiguration is triggered. This secondary reconfiguration is different from the initial scheduling reconfiguration, and its priority strategy is: first, try to adjust the delivery order before the unreachable node or insert additional rest nodes in transit to reduce the predicted physiological load value when arriving at the node; if the threshold still cannot be met, then remove the node from the current path and replace it with an alternative delivery node, while replanning the access order of all subsequent nodes and recalculating the arrival time and corresponding cold chain temperature control parameters of each node.

[0072] By forward extrapolating the three-dimensional coupling attenuation model to future delivery nodes, a shift from passively responding to the current state to proactively predicting future risks is achieved. This enables the scheduling system to identify potentially unreachable nodes before vehicles reach high-altitude or high-physiological-load areas and initiate secondary reconstruction in advance, effectively avoiding delivery failures caused by temporary turnbacks or midway stops. At the same time, the targeted adjustments of the secondary reconstruction reduce frequent changes in the global path, improve the stability and execution efficiency of the scheduling scheme, and significantly reduce the no-show rate and safety risks of cold chain delivery in plateau pastoral areas.

[0073] In a preferred embodiment, inserting rest stops along the way includes: predicting the estimated time and geographical location when the driver will reach a critical state of fatigue based on the changing trend of the driver's physiological load index; selecting service points with oxygen supply conditions as candidate rest stops within a preset range near the estimated geographical location; inserting the candidate rest stops into the delivery route; and recalculating the arrival time and corresponding temperature control parameters of subsequent delivery stops.

[0074] In practice, the dispatch platform records the historical sequence of the driver's physiological load index at fixed time steps (e.g., every 5 minutes), extrapolates its changing trend using linear regression or exponential smoothing, and calculates the estimated time when the load index reaches a preset fatigue threshold (e.g., physiological load index ≥ 0.70) by combining the elevation change rate of the road ahead. Simultaneously, based on the current vehicle location, speed, and route direction, the estimated geographical location corresponding to that time is marked in the geographic information system. Then, using this geographical location as the center and a radius of 5 kilometers as the search range, the pre-stored database of pastoral service points is retrieved, and rest stops, clinics, or nursing stations marked with diffused oxygen supply equipment or portable oxygen cylinders are selected as candidate rest nodes, sorted by proximity and sufficient oxygen supply capacity. After selecting the optimal candidate node, it is inserted into the current delivery route as a new waypoint, and a minimum rest time (e.g., 20 minutes) is set for the driver to recover with oxygen. The access order of subsequent delivery nodes is reordered according to the insertion position. The estimated arrival time of each node is recalculated using the updated three-dimensional coupled attenuation model. Based on the change in the time spent on each segment, the refrigerated area temperature compensation value and refrigeration power adjustment command are generated synchronously and sent to the vehicle cold chain control module for execution.

[0075] By combining physiological load trend prediction with geographical oxygen supply resources, proactive intervention before the fatigue threshold is reached, rather than post-event alarms, effectively preventing safety accidents caused by drivers' altitude sickness. At the same time, the insertion of rest nodes and the linkage adjustment of cold chain temperature control parameters ensure that goods remain in a suitable preservation environment when the transit time is extended, taking into account both driving safety and cold chain quality, and significantly improving the sustainability and reliability of long-distance delivery in plateau pastoral areas.

[0076] In a preferred embodiment, the triggering scheduling reconfiguration in step S3 further includes: when the rate of change of altitude of the road segment ahead is detected to exceed a preset slope threshold, pre-calculating the power attenuation coefficient increment of the road segment ahead; determining whether the vehicle has the ability to climb the slope based on the power attenuation coefficient increment; when the determination result is that it does not have the ability, marking the road segment ahead as a prohibited road segment and planning an alternative route, while recalculating the cold chain temperature control parameters based on the change in the length of the alternative route.

[0077] In practice, the dispatch platform receives real-time GPS and elevation data from the vehicle and calculates the rate of elevation change for the road segment within a preset distance (e.g., 2 kilometers ahead). When this rate of change exceeds a preset gradient threshold (e.g., elevation increase ≥ 60 meters / kilometer, corresponding to approximately a 3.4° gradient), the system inputs the elevation sequence of the road segment ahead, along with the current engine output torque and power, into the power dimension of a three-dimensional coupled attenuation model. It then extrapolates and calculates the incremental power attenuation coefficient at each mileage point on that road segment (i.e., the additional attenuation due to the increased gradient). Finally, considering the vehicle's current load, gear position, and engine rated power, the system estimates the driving torque required for the vehicle to climb at its current speed limit and compares it with the engine's actual available output torque. If the available torque is less than 85% of the required torque, the vehicle is deemed unable to climb the hill. At this point, the system automatically marks the road segment as a restricted area and calls up the pre-stored pastoral road network topology data. Starting from the current vehicle location and ending at the subsequent delivery node, it replans the shortest alternative route to avoid the restricted road segment. At the same time, based on the mileage difference between the alternative route and the original route and the estimated change in travel time, it recalculates the temperature control compensation value according to the cold chain preservation requirements of the goods category and adjusts the set temperature and cooling power of the vehicle-mounted cold chain equipment.

[0078] In a preferred embodiment, the synchronous adjustment of the temperature control parameters of the vehicle-mounted cold chain equipment in step S3 includes: calculating the change in transit time between each delivery node based on the reconstructed scheduling scheme; calculating the temperature control compensation value based on the change in transit time and the cold chain preservation requirements corresponding to the product category; and adjusting the set temperature and cooling power output of the vehicle-mounted cold chain equipment based on the temperature control compensation value.

[0079] In practice, after the scheduling reconfiguration is completed, the scheduling platform extracts the path mileage and estimated travel time between each delivery node before and after the reconfiguration, and calculates the change in transit time ΔT segment by segment (transit time after reconfiguration minus original transit time; a positive value indicates an extension, and a negative value indicates a shortening). Subsequently, based on the category identifier of the goods currently loaded on the vehicle-mounted cold chain equipment, the pre-stored cold chain preservation parameter library is retrieved to obtain the suitable temperature range and the maximum allowable exposure time for that category. For example, refrigerated fresh meat requires -1℃ to 4℃, frozen meat requires -18℃ and below, and raw dairy products are sensitive to temperature fluctuations. Based on the sign and absolute value of ΔT, combined with the sensitivity coefficient of the category to the cumulative effect of temperature, the temperature control compensation value is calculated: if ΔT is positive, the set temperature is reduced proportionally and the cooling power is increased to offset the temperature rise effect caused by the cumulative heat exchange inside and outside the container; if ΔT is negative, the set temperature is appropriately increased and the cooling power is reduced while ensuring that the goods are not too cold, in order to save energy. The generated compensation command is sent to the vehicle-mounted cold chain control module via the CAN bus. The latter adjusts the compressor unit speed and expansion valve opening to stabilize the temperature inside the chamber at the compensated target value.

[0080] By mapping the changes in transit time caused by scheduling reconfiguration to quantitative adjustments of cold chain temperature control parameters in real time, the traditional static operation mode of cold chain temperature control, which is independent of scheduling decisions, has been changed. This has enabled the coordinated optimization of transportation plan adjustments and cargo preservation requirements. Differentiated compensation for different product categories avoids overcooling damage or energy waste caused by one-size-fits-all adjustments, effectively ensuring the consistency of cargo quality in cold chain logistics in plateau pastoral areas under dynamic scheduling.

[0081] In a preferred embodiment, step S4 further includes: recording the triggering conditions for each scheduling reconfiguration, the path differences before and after reconfiguration, and the actual execution deviation; training a reinforcement learning model based on the triggering conditions, path differences, and actual execution deviation; using the trained reinforcement learning model to predict the scheduling risk probability of future road segments, and triggering pre-scheduling reconfiguration in advance when the scheduling risk probability exceeds a preset threshold.

[0082] In practice, each scheduling reconfiguration event is used as a sample unit to record the triggering conditions (such as blood oxygen saturation below the first threshold, power attenuation coefficient above the second threshold, and slope prediction failure), the differences in delivery node sequences and path mileage before and after reconfiguration, and the deviation values ​​(including the number of minutes of early or late arrival) between the actual arrival time of each node and the planned time. The above data is stored in a historical case database by timestamp and extracted as feature vectors, including the current altitude change rate, the slope of the physiological load index change, the remaining power attenuation coefficient, the average slope of the road segment, and the weather level. A deep reinforcement learning network (such as the DQN architecture) is used for offline training, with the historical feature sequence as input and the probability of triggering reconfiguration within a preset future time period as output. The reward function is set to the negative correlation value of the actual deviation (the smaller the deviation, the higher the reward). After training convergence, the model is deployed to the scheduling platform. During the driving process, the current perception data is input in a scrolling window (e.g., every 15 minutes), and the risk probability value of each road segment is output for the next 2 hours. When the risk probability of any road segment exceeds the preset threshold (e.g., ≥0.75), the system triggers pre-scheduling reconstruction in advance. That is, before the risk actually occurs, the system actively performs operations such as path correction, speed limit adjustment, or pre-insertion of rest nodes.

[0083] Through the accumulation of historical reconstruction cases and reinforcement learning training, the system can identify the precursor patterns of scheduling failures in complex high-altitude scenarios from experience, transforming passive response into proactive prevention, effectively reducing the frequency of emergency reconstruction and execution deviations. At the same time, pre-schedule reconstruction provides drivers and cold chain equipment with more buffer time, reducing the impact of sudden adjustments on transportation quality, and significantly improving the system's long-term operational stability and adaptability in uncertain environments such as weak communication and high altitude.

[0084] This embodiment also provides a dynamic scheduling system for logistics vehicles, implementing a dynamic scheduling method for logistics vehicles as described in the above embodiment. The system includes a multi-source perception layer, a model building layer, a scheduling decision layer, a closed-loop optimization layer, and an on-board cold chain control module. The multi-source perception layer includes on-board environmental sensors, a driver wearable device, and a vehicle CAN bus interface, used to collect real-time data on altitude, atmospheric pressure, ambient oxygen content, blood oxygen saturation, heart rate, engine output power, and output torque. The model building layer is used to calculate the vehicle power attenuation coefficient and the driver's physiological load index based on the data collected by the multi-source perception layer, and to construct a three-dimensional coupled attenuation system based on the environment, physiology, and power. The system consists of a three-dimensional coupled attenuation model and a scheduling decision layer. The scheduling decision layer assesses the accessibility of each delivery node on the preset delivery route based on the three-dimensional coupled attenuation model. When the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold, the system outputs the reconstructed delivery node sequence, the maximum speed limit of the road segment, the rest nodes along the way, and the delivery time window. At the same time, it generates cold chain temperature control parameter adjustment instructions. The closed-loop optimization layer continuously receives data from the multi-source sensing layer during the driving process, updates the three-dimensional coupled attenuation model, and iteratively optimizes the scheduling decision. The on-board cold chain control module receives cold chain temperature control parameter adjustment instructions and executes adjustments to the set temperature and cooling power output.

[0085] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic scheduling of logistics vehicles, characterized in that, Includes the following steps: S1. The vehicle uses onboard environmental sensors to obtain the current road section's altitude, atmospheric pressure, and ambient oxygen content in real time; wearable devices can be used to obtain the driver's blood oxygen saturation and heart rate in real time; and the vehicle's CAN bus can be used to obtain the engine's output power and output torque in real time. S2. Based on the altitude, atmospheric pressure and ambient oxygen content, calculate the vehicle power attenuation coefficient and the driver's physiological load index respectively. Based on the vehicle power attenuation coefficient and the driver's physiological load index, construct a three-dimensional coupled attenuation model to characterize the linkage attenuation relationship between the vehicle's actual carrying capacity and the driver's continuous working ability. S3. Based on the three-dimensional coupled attenuation model, the accessibility of each delivery node on the preset delivery route is evaluated. When the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold, the scheduling reconstruction is triggered. The scheduling reconstruction includes replanning the order of delivery nodes, adjusting the upper limit of the road segment driving speed, inserting rest nodes in the road and correcting the delivery time window. According to the reconstructed scheduling scheme, the temperature control parameters of the vehicle cold chain equipment are adjusted synchronously to compensate for the change in the in-transit time. S4. During the driving process, the multi-source sensing data is continuously collected, the three-dimensional coupling attenuation model is updated, and the scheduling decision is iteratively optimized based on the updated model until all delivery tasks are completed.

2. The dynamic scheduling method for logistics vehicles according to claim 1, characterized in that, The calculation of the vehicle power attenuation coefficient includes: The baseline attenuation rate is obtained by querying the pre-stored dynamic attenuation mapping table based on the altitude. The baseline attenuation rate is adjusted based on the atmospheric pressure and ambient oxygen content to obtain the vehicle power attenuation coefficient.

3. The dynamic scheduling method for logistics vehicles according to claim 1, characterized in that, The calculation of the driver's physiological workload index includes: The blood oxygen saturation and heart rate data are normalized to obtain normalized blood oxygen values ​​and normalized heart rate values. The blood oxygen weighting coefficient and heart rate weighting coefficient are determined based on the altitude. The driver's physiological load index is calculated based on a weighted summation.

4. The dynamic scheduling method for logistics vehicles according to claim 1, characterized in that, The accessibility assessment includes: Based on the current vehicle location, delivery node sequence, and maximum speed limit of the road section, the estimated arrival time of the vehicle at each delivery node is determined. Based on the three-dimensional coupled attenuation model, the predicted values ​​of the power attenuation coefficient and the physiological load index are estimated when the vehicle reaches each delivery node. When the predicted value exceeds the preset safety threshold, the corresponding delivery node is marked as an unreachable node, and a secondary scheduling reconfiguration is triggered.

5. The dynamic scheduling method for logistics vehicles according to claim 1, characterized in that, The insertion of the in-transit rest node includes: Based on the changing trend of the driver's physiological load index, the estimated time and geographical location when the driver reaches the fatigue threshold are predicted. Within a preset range near the estimated geographical location, service points with oxygen supply conditions are selected as candidate rest nodes. The candidate rest nodes are inserted into the delivery path, and the arrival times and corresponding temperature control parameters of subsequent delivery nodes are recalculated.

6. The method for dynamic scheduling of logistics vehicles according to claim 1, characterized in that, The synchronous adjustment of the temperature control parameters of the vehicle-mounted cold chain equipment includes: Calculate the change in transit time between each delivery node based on the reconstructed scheduling scheme; Calculate the temperature control compensation value based on the change in transit time and the cold chain preservation requirements corresponding to the product category; The set temperature and cooling power output of the vehicle-mounted cold chain equipment are adjusted according to the temperature control compensation value.

7. The dynamic scheduling method for logistics vehicles according to claim 1, characterized in that, In step S3, the triggered scheduling reconfiguration further includes: When the rate of change of altitude of the road section ahead is detected to exceed the preset slope threshold, the increment of the dynamic attenuation coefficient of the road section ahead is pre-calculated. The vehicle's ability to climb hills is determined based on the increase in the power attenuation coefficient. When the judgment result is not met, the road section ahead is marked as a no-entry section, and an alternative route is planned. At the same time, the cold chain temperature control parameters are recalculated based on the change in the length of the alternative route.

8. The method for dynamic scheduling of logistics vehicles according to claim 1, characterized in that, Step S4 further includes: Record the triggering conditions for each scheduling refactoring, the path differences before and after the refactoring, and the actual execution deviation; The reinforcement learning model is trained based on the aforementioned triggering conditions, path differences, and actual execution deviations. The trained reinforcement learning model is used to predict the scheduling risk probability of future road segments, and pre-scheduling reconstruction is triggered in advance when the scheduling risk probability exceeds a preset threshold.

9. The dynamic scheduling method for logistics vehicles according to claim 2, characterized in that, Step S1 also includes: The satellite communication module is used to obtain real-time road conditions and weather warnings for each delivery node within a pre-defined area of ​​the pastoral region. The real-time road traffic status and weather warning information are input as external constraints into the three-dimensional coupled attenuation model to correct the accessibility assessment results.

10. A dynamic scheduling system for logistics vehicles, implementing the dynamic scheduling method for logistics vehicles as described in any one of claims 1-9, characterized in that, The system includes: The multi-source sensing layer includes onboard environmental sensors, driver wearable devices, and the vehicle CAN bus interface, used to collect data in real time, such as altitude, atmospheric pressure, ambient oxygen content, blood oxygen saturation, heart rate, engine output power, and output torque. The model building layer is used to calculate the vehicle power attenuation coefficient and the driver's physiological load index based on the data collected by the multi-source perception layer, and to build a three-dimensional coupled attenuation model. The scheduling decision layer is used to assess the accessibility of each delivery node on the preset delivery route based on the three-dimensional coupled attenuation model, and triggers scheduling reconstruction when the blood oxygen saturation is lower than the preset first threshold or the power attenuation coefficient is higher than the preset second threshold. It outputs the reconstructed delivery node sequence, the upper limit of the road segment driving speed, the rest nodes in the way and the delivery time window, and generates cold chain temperature control parameter adjustment instructions. A closed-loop optimization layer is used to continuously receive data from the multi-source perception layer during driving, update the three-dimensional coupling attenuation model, and iteratively optimize scheduling decisions. The vehicle-mounted cold chain control module is used to receive the cold chain temperature control parameter adjustment command and execute the adjustment of the set temperature and cooling power output.

Images