Intelligent traffic charging data consensus system based on block chain

Abstract

According to the intelligent traffic charging data consensus system based on the block chain, a dynamically optimized hybrid consensus mechanism is adopted, a node election algorithm and a fragment consensus strategy are designed in combination with traffic flow features, and high concurrent processing capacity is achieved. The system ensures non-tampering and traceability of charging data through technologies such as credible data acquisition, intelligent contract automatic execution, hierarchical storage architecture and the like, and meets the real-time supervision requirement at the same time. A specially designed privacy protection module protects sensitive information of a user on the premise of ensuring supervision transparency, and a proposed cross-chain reconciliation protocol effectively solves the problem of multi-region collaborative charging.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic charging systems, and in particular relates to a blockchain-based intelligent traffic data consensus system and a method for using the system. Background Art

[0002] With the rapid development of intelligent transportation systems, the existing transportation charging systems have gradually exposed the following problems in terms of technical architecture and operation mode: 1. Vulnerability of centralized architecture: Traditional toll collection systems often utilize centralized data management centers, such as provincial toll settlement platforms. This architecture presents a single point of failure, as evidenced by a 12-hour outage at all toll stations in a province's expressway toll collection system in 2019 caused by a downtime in the primary database server. Furthermore, centralized traffic data is susceptible to malicious tampering. An audit revealed that between 2018 and 2020, a city in the province experienced 23,000 disputes involving abnormal toll deductions due to system vulnerabilities, involving over 10 million yuan.

[0003] 2. Dilemma of cross-platform data interaction: Because regional toll collection systems utilize heterogeneous technology standards (for example, ETC utilizes 5.8GHz DSRC communication, while some cities use RFID identification), vehicles traveling across provinces and cities must manually verify their data through a reconciliation platform. Statistics show that the error rate for interprovincial transactions is as high as 0.7%, and the average settlement cycle is as long as 72 hours, severely hindering the achievement of the goal of a nationwide "single network" operation.

[0004] 3. The contradiction between privacy protection and regulatory transparency: Existing systems transmit vehicle trajectory information in plain text during data collection (for example, one ETC vendor's data package contains unencrypted OBU ID and GPS coordinates), posing a risk of user privacy leakage. Furthermore, the use of traditional encryption methods (such as SSL / TLS) prevents regulators from effectively tracing abnormal transactions, creating a dilemma: either compromise privacy or compromise auditability.

[0005] 4. Lag of the dynamic pricing mechanism: Current congestion charging systems rely on fixed-period pricing strategies (such as Beijing's morning rush hour zone charging scheme) and are unable to respond to real-time changes in traffic flow. 2021 field data from the Shanghai Inner Ring Elevated Road showed that traditional systems experience a 15-30 minute delay from traffic monitoring to rate adjustments, reducing the effectiveness of peak-hour traffic control by over 40%.

[0006] 5. Adaptability issues of blockchain applications: Although there have been attempts to apply blockchain technology to the toll collection field, the single consensus mechanism it adopts (such as PoW or ordinary PBFT) is difficult to adapt to the characteristics of traffic scenarios: the number of ordinary PBFT nodes is fixed and cannot cope with the surge in traffic during holidays (for example, the number of toll station nodes needs to be expanded by 300% during the Spring Festival travel rush); traditional DPoS lacks a trusted election mechanism. In 2019, a malicious node in an experimental system manipulated the rate calculation by forging reputation values. The spatiotemporal correlation of traffic data is not taken into account, resulting in inefficient consensus verification (data of the same vehicle at adjacent toll stations needs to be verified across shards).

[0007] To cater to the development of the intelligent era and the increasing traffic data throughput, the traditional intelligent toll collection system functions can no longer meet the current processing needs, and a more intelligent and efficient processing system is needed to cope with big data. Summary of the Invention

[0008] In response to the above-mentioned defects, the present invention provides a blockchain-based intelligent transportation charging data consensus system, which solves the defects of traditional systems in terms of security, real-time performance, scalability, etc., and provides a reliable, efficient and compliant new infrastructure for intelligent transportation charging.

[0009] The present invention is achieved through the following technical solutions: A blockchain-based intelligent traffic toll collection data consensus system, including a data collection layer, a blockchain network layer, a smart contract layer, an application service layer, and a consensus engine; The data collection layer is deployed in the IoT node network of toll booths, ETC equipment and on-board OBU terminals. Each node integrates a trusted execution environment (TEE) chip, encrypts the collected vehicle ID, travel time, and location coordinates using the national secret SM4 algorithm, and generates a data packet containing the device digital certificate. The blockchain network layer consists of a two-layer architecture consisting of a regional main chain and multiple road segment sub-chains. The main chain nodes are composed of traffic management agency servers, and the sub-chain nodes include roadside units (RSUs) and edge computing gateways. The network layer has a built-in dynamic sharding module, which divides the network into multiple shards based on real-time traffic flow. Each shard contains at least 11 consensus nodes. The smart contract layer is deployed in the virtual machine of the blockchain network layer and includes the following contracts: Rate calculation contract: uses the LSTM prediction model to calculate dynamic rate coefficients based on real-time traffic flow data; Cross-chain reconciliation contract: uses the Hash Time Lock Protocol (HTLC) to achieve atomic fund exchange between different sub-chains; Data verification contract: Verify the validity of encrypted data through zero-knowledge proof without exposing the original information; The application service layer provides a data sharing API interface based on the OAuth2.0 protocol and a supervision platform integrated with the ECharts visualization engine; The consensus engine adopts an improved DPoS-PBFT hybrid mechanism, in which the DPoS stage elects witness nodes based on the node reputation value range of 0-100 and the traffic flow weight of the shard to which it belongs; the PBFT stage adopts a three-phase commit optimization protocol within the shard, adds data spatiotemporal feature verification in the pre-preparation stage, and introduces abnormal data detection based on cosine similarity in the confirmation stage.

[0010] This paper designs a dynamic sharding-optimized DPoS-PBFT hybrid consensus mechanism, which reduces the cross-regional transaction confirmation time from seconds to sub-seconds while ensuring Byzantine fault tolerance. It also innovates a multi-dimensional node reputation assessment model, combining multi-source data such as device security authentication, historical behavior analysis, and network quality monitoring to increase the accuracy of malicious node identification to 99.2%. It develops a smart contract layer that supports privacy computing, adopts zero-knowledge proof + homomorphic encryption dual modes, and realizes a "data available but invisible" regulatory compliance solution. It constructs a dynamic rate model that adapts to traffic flow, and through closed-loop control of LSTM prediction and real-time consensus feedback, compresses the rate adjustment delay to less than 5 minutes. It develops a hierarchical verifiable storage architecture to greatly improve the efficiency of historical transaction queries while ensuring data integrity.

[0011] As a further improvement of the present invention, the specific implementation of the dynamic sharding module includes: a) Slice generation unit: Calculates the road segment association every 5 minutes using a road network topology analysis algorithm and groups adjacent toll booth nodes into the same slicing using a spectral clustering method; b) Traffic prediction unit: uses the ARIMA model to predict the traffic volume within the shard for the next 15 minutes. When the predicted value exceeds the threshold Q = 100 vehicles / minute, shard fission is automatically triggered; c) Node Scheduling Unit: Dynamically adjust the node composition within the shard based on the historical consensus success rate and network latency <200ms. The scheduling strategy satisfies the Byzantine fault tolerance condition 3f+1≤N.

[0012] As a further improvement of the present invention, the calculation process of the node reputation value is as follows: The input dimensions include: the device firmware hash value matching degree α has a weight of 20%, the number of data objections in the past 30 days β has a weight of 35%, the network heartbeat packet loss rate γ has a weight of 15%, and the block height continuity δ participating in the consensus has a weight of 30%. The weight of each dimension is calculated using the entropy weight method, and the reputation value R=Σ(α_i×w_i)×100. When R<60, it is marked as an abnormal node; The reputation ranking list is updated through the alliance chain governance contract at 00:00 every day, and the top 20% of nodes are qualified as witness node candidates.

[0013] As a further improvement of the present invention, the encryption processing of the data acquisition layer includes: Perform segmented encryption in the TEE environment: vehicle identity information is encrypted using SM4-CTR mode, and location data uses the homomorphic encryption Paillier algorithm; Data fingerprint generation: Perform two SHA-3 hash operations on the encrypted data packet, and concatenate the first hash result with the device GPS timestamp and perform a second hash. The edge computing gateway performs data cleaning: it eliminates abnormal records with speeds greater than 220 km / h or with the same vehicle passing through adjacent toll stations with a time interval less than Δt; Δt = distance between the two stations / maximum speed limit.

[0014] As a further improvement of the present invention, the execution logic of the rate calculation contract is: The basic rate table is preset in the contract storage area, including time period coefficients of 1.2 for morning peak, 1.0 for off-peak, and 0.8 for night, and vehicle type coefficients of 1.0 for small cars and 2.5 for large cars. The dynamic adjustment module obtains real-time data from traffic flow sensors every 15 minutes. When the road section saturation K = actual flow / design capacity > 0.9, the rate is triggered to increase by ΔP = base price × (K-0.9) × adjustment coefficient η, where η = 0.05-0.2, a configurable parameter. Cost calculation formula: Total cost = base price × mileage × (time period coefficient + dynamic adjustment coefficient) × vehicle model coefficient.

[0015] As a further improvement of the present invention, the visual supervision platform includes: Real-time monitoring panel: Renders the provincial road network status in GeoJSON format. The color gradient indicates the level of traffic congestion. Green / yellow / red correspond to V / C ratios of 0-0.6 / 0.6-0.9 / 0.9+; Audit tracking module: Entering a license plate number can trace its complete traffic record, displaying a three-level evidence chain including block height, consensus node signature, and storage location CID; Risk warning system: triggers an alarm when the following situations are detected: i) A single node has three consecutive abnormal voting blocks ii) The rate for the same road section fluctuates by more than 15% within 1 hour iii) Cross-chain transaction confirmation time > 5 seconds Regulatory sandbox: supports importing simulated data to test contract logic and provides a transaction rollback testing interface.

[0016] As a further improvement of the present invention, the specific consensus process of the DPoS-PBFT hybrid mechanism is as follows: Phase 1 Witness Node Election: a) Each shard selects the top N = floor (total number of nodes × 15%) candidate nodes based on reputation ranking; b) The candidate node submits a campaign proposal containing the hash of the most recent block; c) The final 21 witness nodes are determined through the VRF random number generator; Phase 2 transaction packaging: d) The witness node collects transactions within the shard and generates a new block proposal containing the Merkle root, the previous block hash, and the time window signature; Phase 3: PBFT consensus within shards: e) Pre-preparation phase: The master node broadcasts the block header with RSA signature to all nodes in the shard; f) Preparation phase: Verification nodes check timestamp validity and double-spending transactions, with a deviation of <±5s; g) Confirmation phase: Each node performs threshold signatures on the block, and confirmation is completed when 2f+1 valid signatures are collected; Phase 4 cross-shard synchronization: h) Broadcast the block header to other shards through the Gossip protocol and perform light node verification.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The system of the present invention adopts the DPoS-PBFT hybrid consensus mechanism and dynamic sharding technology, which greatly improves the system throughput and can automatically adjust the sharding scale according to traffic flow fluctuations, solving the core problems of blockchain technology in traffic scenarios, such as insufficient scalability and poor real-time response.

[0018] 2. The invented system adopts a dual-mode verification architecture of zero-knowledge proof + homomorphic encryption to complete data verification in an encrypted state, reducing the risk of sensitive information leakage by 98%. At the same time, a multi-level audit tracking chain of block height + node signature + IPFS storage proof ensures penetrable audits by regulatory agencies, solving the dilemma of "privacy protection" and "regulatory transparency".

[0019] 3. The system of the present invention is based on a dynamic rate system with an LSTM prediction model and a consensus feedback closed loop, which reduces the rate adjustment delay and realizes multi-region rate coordination through a cross-chain atomic transaction protocol, significantly improving traffic diversion efficiency.

[0020] 4. The system of the present invention constructs a hierarchical verifiable storage architecture, which makes the response time of the hot data layer less than 50ms and improves the efficiency of cold data retrieval. Combined with a multi-dimensional node reputation evaluation model with high accuracy in identifying malicious nodes, the data tampering detection time is greatly shortened, thereby reducing the system operation and maintenance costs. DETAILED DESCRIPTION

[0021] The present invention will be further described below with reference to the examples. In the examples, unless otherwise specified, the technical means used are all conventional technical means in the art. Example 1

[0022] This system, deployed in a demonstration area for integrated transportation, includes multimodal transportation fee settlement, including ETC, mobile payment, and truck weight-based toll collection, while also implementing dynamic congestion pricing. The network encompasses 15 test highways and 83 test toll booths, processing an average of 2 million transactions per day, with peak times reaching 5,000 transactions per second.

[0023] Hardware configuration: The industrial computers in the toll station lanes are equipped with the Intel SGX trusted execution environment; the OBU terminals are integrated with the national encryption SM2 / 4 chips; and the roadside edge computing nodes are equipped with NVIDIA Jetson Xavier computing modules.

[0024] Data stream processing: When vehicle license plate number Beijing A12345 passes through test toll booth 1 on test highway 1, the OBU collects the license plate (ciphertext 0x7A3D...F9C1), axle weight (12.3 tons → homomorphically encrypted value), and travel time. The laser scanner obtains vehicle model features, and data fusion is performed within the TEE. SM4-CTR encryption (key index number 20231001-085) is applied to the original data packet to generate a 128-byte encrypted data block.

[0025] Edge node preprocessing: Perform spatiotemporal validity checks: Compare data from adjacent test toll booths to verify that the travel time difference is ≥ ΔT = 18 km between the two booths / 60 km / h minimum speed limit = 1080 seconds. Generate a data fingerprint using the SHA3-512 (encrypted data || device certificate SN: HBWY-ETC-2309-0875) command. Output a 64-byte hash value of 0x8E3A...D7F2.

[0026] Network topology construction: Main chain nodes: deployed in the transportation data centers in locations A, B, and C, each equipped with three redundant servers; Subchain sharding: Based on the morning peak traffic volume from 08:00 to 09:00, the road network is divided into 5 shards: Segment 1: Test Highway 1-Test Highway 2 (12 toll booths, predicted traffic flow Q = 142 vehicles / minute); Segment 2: Test Highway 3-Test Highway 4 (9 toll booths, Q=98 vehicles / minute); ...(the rest of the fragments are omitted).

[0027] Consensus mechanism execution: 1. Node election phase: 38 nodes in shard 1 participate in the election, and the reputation evaluation module calculates the score of each node: Node BJ-N025: Firmware hash match 100% (weight 20% → 20 points); 0 objections in the past 30 days (weight 35% → 35 points); network latency 183ms (weight 15% → 13.05 points); 2000 consecutive block participations (weight 30% → 27 points); total reputation R = 20 + 35 + 13.05 + 27 = 95.05; the top 15% of nodes (6) are selected as witness candidates, and 4 witness nodes are randomly selected through VRF.

[0028] 2. Transaction packaging and verification: Witness node BJ-N025 collected 350 transactions and constructed block #589032: the block header contains the parent hash 0x3A7D...CF21 and the timestamp 2023-10-01T08:16:00Z; the Merkle tree is constructed using a modified Patricia Trie structure with a depth of 6 layers.

[0029] Optimized PBFT execution within shards: During the pre-preparation phase, the time and space constraints are verified (vehicle Beijing A12345 passes through test toll station 1 at 08:15 and arrives at test toll station 2 at 08:33 → speed = 18 km / 0.3 h = 60 km / h, which meets the speed limit); During the preparation phase, 11 validators check for double-spending transactions and use Bloom filters for fast matching. In the confirmation phase, consensus is reached after collecting 8 valid signatures (f=3, satisfying 2f+1=7).

[0030] 3. Cross-shard synchronization: The block header is broadcast to other shards through the improved Gossip protocol, and Merkle proof is used to implement light node verification, which takes only 72ms.

[0031] 1. Dynamic rate calculation: LSTM model input parameters: Real-time traffic: 386 vehicles / 15 minutes at the entrance of toll station 1; Weather data: Haze visibility 800 meters (weighting coefficient +0.15); Historical average for the same period: 352 vehicles / 15 minutes; Forecast output: Traffic growth rate ΔQ = +18% in the next 30 minutes.

[0032] 2. The rate adjustment smart contract executes, and the output result is: Beijing A12345 truck toll = 0.5 ETH / km × 18km × (1+0.15) × 2.5 = 25.875 ETH.

[0033] 3. Cross-chain settlement When the vehicle enters Section 5 of the test highway, the HTLC protocol is triggered: Subchain A locks 25.875 ETH and generates a hash lock H=keccak256(secret). Subchain B submits the preimage within 300 seconds to complete the atomic exchange of funds. Subchain C acts as a notary to verify the transaction and charges a 0.1% handling fee.

[0034] Traffic Management Bureau logs into the visualization platform: Real-time monitoring: The V / C ratio of the high-speed section 1 was found to be 0.92 (red alert); Penetrating audit: Query the full chain data of Beijing A12345: Data source: Signature verification by 3 edge nodes.

[0035] Consensus proof: BLS signatures from 8 nodes in shard 1.

[0036] Proof of Storage: IPFS CID verifiability check passed.

[0037] Risk early warning system triggers: It was detected that node TJ-N112 of shard 2 had abnormal voting for three consecutive blocks.

[0038] Automatically start the node isolation program and trigger the reputation reassessment.

[0039] This implementation achieves a peak transaction throughput of 5032 TPS, with an average confirmation time of 0.9 seconds. Atomic transactions with cross-provincial settlements achieve a 99.98% success rate, with an average transaction time of 1.5 seconds. Zero-knowledge verification takes less than 200ms, and the key cracking difficulty reaches 2^128. Compared to traditional systems, the risk of data tampering is significantly reduced, reconciliation labor costs are reduced, network sharding efficiency is improved, and storage costs are reduced.

[0040] The above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the scope of the present invention. The scope of protection of the present invention is defined by the claims. Persons skilled in the art may make various modifications or equivalent substitutions to the present invention within the spirit and scope of protection of the present invention, and such modifications or equivalent substitutions shall also be deemed to fall within the scope of protection of the present invention.

Claims

1. A blockchain-based intelligent traffic charging data consensus system, characterized by: Includes data collection layer, blockchain network layer, smart contract layer, application service layer and consensus engine; The data collection layer is deployed in the IoT node network of toll booths, ETC equipment and on-board OBU terminals. Each node integrates a trusted execution environment (TEE) chip, encrypts the collected vehicle ID, travel time, and location coordinates using the national secret SM4 algorithm, and generates a data packet containing the device digital certificate. The blockchain network layer consists of a two-layer architecture consisting of a regional main chain and multiple road segment sub-chains. The main chain nodes are composed of traffic management agency servers, and the sub-chain nodes include roadside units (RSUs) and edge computing gateways. The network layer has a built-in dynamic sharding module, which divides the network into multiple shards based on real-time traffic flow. Each shard contains at least 11 consensus nodes. The smart contract layer is deployed in the virtual machine of the blockchain network layer and includes the following contracts: Rate calculation contract: uses the LSTM prediction model to calculate dynamic rate coefficients based on real-time traffic flow data; Cross-chain reconciliation contract: uses the Hash Time Lock Protocol (HTLC) to achieve atomic fund exchange between different sub-chains; Data verification contract: Verify the validity of encrypted data through zero-knowledge proof without exposing the original information; The application service layer provides a data sharing API interface based on the OAuth2.0 protocol and a supervision platform integrated with the ECharts visualization engine; The consensus engine adopts an improved DPoS-PBFT hybrid mechanism, in which witness nodes are elected during the DPoS phase based on the node reputation value range of 0-100 and the traffic weight of the shard to which it belongs; In the PBFT phase, a three-phase commit optimization protocol is adopted within the shards. Data temporal and spatial feature verification is added to the pre-preparation phase, and abnormal data detection based on cosine similarity is introduced in the confirmation phase.

2. The blockchain-based intelligent transportation charging data consensus system according to claim 1 is characterized by: The specific implementation of the dynamic sharding module includes: a) Slice generation unit: Calculates the road segment association every 5 minutes using a road network topology analysis algorithm and groups adjacent toll booth nodes into the same slicing using a spectral clustering method; b) Traffic prediction unit: uses the ARIMA model to predict the traffic volume within the shard for the next 15 minutes. When the predicted value exceeds the threshold Q = 100 vehicles / minute, shard fission is automatically triggered; c) Node Scheduling Unit: Dynamically adjust the node composition within the shard based on the historical consensus success rate and network latency <200ms. The scheduling strategy satisfies the Byzantine fault tolerance condition 3f+1≤N.

3. The blockchain-based intelligent transportation charging data consensus system according to claim 1 is characterized by: The calculation process of the node reputation value is as follows: The input dimensions include: the device firmware hash value matching degree α has a weight of 20%, the number of data objections in the past 30 days β has a weight of 35%, the network heartbeat packet loss rate γ has a weight of 15%, and the block height continuity δ participating in the consensus has a weight of 30%. The weight of each dimension is calculated using the entropy weight method, and the reputation value R=Σ(α_i×w_i)×100. When R<60, it is marked as an abnormal node; The reputation ranking list is updated through the alliance chain governance contract at 00:00 every day, and the top 20% of nodes are qualified as witness node candidates.

4. The blockchain-based intelligent transportation charging data consensus system according to claim 1 is characterized by: The encryption processing of the data acquisition layer includes: Perform segmented encryption in the TEE environment: vehicle identity information is encrypted using SM4-CTR mode, and location data uses the homomorphic encryption Paillier algorithm; Data fingerprint generation: Perform two SHA-3 hash operations on the encrypted data packet, and concatenate the first hash result with the device GPS timestamp and perform a second hash. The edge computing gateway performs data cleaning: it eliminates abnormal records with speeds greater than 220 km / h or with the same vehicle passing through adjacent toll stations with a time interval less than Δt; Δt = distance between the two stations / maximum speed limit.

5. The blockchain-based intelligent transportation charging data consensus system according to claim 1 is characterized by: The execution logic of the rate calculation contract is: The basic rate table is preset in the contract storage area, including time period coefficients of 1.2 for morning peak, 1.0 for off-peak, and 0.8 for night, and vehicle type coefficients of 1.0 for small cars and 2.5 for large cars. The dynamic adjustment module obtains real-time data from traffic flow sensors every 15 minutes. When the road section saturation K = actual flow / design capacity > 0.9, the rate is triggered to increase by ΔP = base price × (K-0.9) × adjustment coefficient η, where η = 0.05-0.2, a configurable parameter. Cost calculation formula: Total cost = base price × mileage × (time period coefficient + dynamic adjustment coefficient) × vehicle model coefficient.

6. The blockchain-based intelligent transportation charging data consensus system according to claim 1 is characterized by: The visual supervision platform includes: Real-time monitoring panel: Renders the provincial road network status in GeoJSON format. The color gradient indicates the level of traffic congestion. Green / yellow / red correspond to V / C ratios of 0-0.6 / 0.6-0.9 / 0.9+; Audit tracking module: Entering a license plate number can trace its complete traffic record, displaying a three-level evidence chain including block height, consensus node signature, and storage location CID; Risk warning system: triggers an alarm when the following situations are detected: i) A single node has three consecutive abnormal voting blocks ii) The rate for the same road section fluctuates by more than 15% within 1 hour iii) Cross-chain transaction confirmation time > 5 seconds Regulatory sandbox: supports importing simulated data to test contract logic and provides a transaction rollback testing interface.

7. The blockchain-based intelligent transportation charging data consensus system according to claim 1 is characterized by: The specific consensus process of the DPoS-PBFT hybrid mechanism is as follows: Phase 1 Witness Node Election: a) Each shard selects the top N = floor (total number of nodes × 15%) candidate nodes based on reputation ranking; b) The candidate node submits a campaign proposal containing the hash of the most recent block; c) The final 21 witness nodes are determined through the VRF random number generator; Phase 2 transaction packaging: d) The witness node collects transactions within the shard and generates a new block proposal containing the Merkle root, the previous block hash, and the time window signature; Phase 3: PBFT consensus within shards: e) Pre-preparation phase: The master node broadcasts the block header with RSA signature to all nodes in the shard; f) Preparation phase: Verification nodes check timestamp validity and double-spending transactions, with a deviation of <±5s; g) Confirmation phase: Each node performs threshold signatures on the block, and confirmation is completed when 2f+1 valid signatures are collected; Phase 4 cross-shard synchronization: h) Broadcast the block header to other shards through the Gossip protocol and perform light node verification.