China’s Transport Ministry Keeps ETC Devices as License-Plate Payment Moves Forward

China’s Transport Ministry has rejected an immediate ETC device phaseout, despite preparing a license-plate payment option for drivers without onboard equipment.

The response followed a motorist’s proposal to connect a vehicle’s plate with Alipay, WeChat Pay, or a bank account. Cameras could then identify the car and collect its toll without a separate device.

That proposal sounds like a simple hardware upgrade. The ministry’s answer shows why it is actually a choice between two different operating models.

ETC uses a radio link between roadside equipment and a vehicle-mounted transponder. The alternative depends primarily on cameras, optical character recognition, payment accounts, and back-office matching.

China will begin expanding that alternative, but it will initially supplement ETC rather than replace it. The first deployments will also retain stopping at manual toll lanes.

The distinction matters across China’s 199,400-kilometer expressway network. A payment method that performs well during a small trial must also survive weather, plate damage, account failures, and holiday traffic nationwide.

License-plate payment is technically feasible. Toll roads in Canada, Britain, and other markets already use versions of it.

The harder question is whether cameras alone can match ETC’s combination of speed, identification certainty, pricing accuracy, and enforcement. China’s current policy says they cannot yet do so at national scale.

The Ministry Approved a Limited Form of License-Plate Payment

China is not choosing between ETC and license-plate payment yet. It is assigning each method a different role.

The underlying event became clear on August 29, 2026. A published ministry response described how the Transport Ministry’s Highway Bureau answered the motorist.

The driver argued that ETC installation consumes time and resources. The proposal would bind a plate directly to a digital wallet or bank account.

According to the bureau, ETC remains an important tool for improving toll-road throughput. Drivers can still decide voluntarily whether to obtain an onboard device.

However, the response also disclosed a national implementation path for drivers without one. The ministry has issued a 2026 to 2028 plan for cardless travel using smartphones and license-plate recognition.

Under that model, cameras identify the vehicle at toll-road control points. The system reconstructs its route, calculates the charge, and sends the driver to an electronic payment process.

Eligible cars would no longer collect a CPC card when entering. A CPC card is the physical toll card used to record a non-ETC vehicle’s route.

Removing that exchange eliminates two manual actions. Drivers would not need to take a card at entry or return it at exit.

The initial experience still differs from true ETC travel. Drivers will use existing MTC lanes, meaning lanes that traditionally handle manual toll collection.

The ministry describes the process as cardless, not completely barrierless. Payment still occurs through the electronic options available in those lanes.

By the end of 2026, the ministry expects coverage for eligible passenger vehicles across selected MTC lanes. The stated scope includes the Chengdu-Chongqing region and the Yangtze River Delta.

The plan also covers MTC exits across the broader network. Eligibility initially applies to small passenger vehicles with no more than nine seats.

Expansion into ETC-only lanes will depend on trial results. This sequence turns the proposal into a measured operational test instead of a nationwide replacement order.

That difference often disappears in abbreviated discussions of the announcement. The ministry did not say license-plate payment was impossible.

It also did not agree that a camera and payment account can immediately replace every onboard unit. Its answer created a second access path while preserving ETC.

For occasional highway users, that second path removes installation friction. It also reduces dependence on physical cards and staffed exchanges.

Frequent travelers still receive the faster option through ETC. The radio transaction can identify an account before a camera-only workflow finishes its back-office checks.

The policy therefore separates convenience from maximum throughput. License-plate recognition improves access, while the ETC device remains the preferred high-frequency credential.

That compromise reflects China’s existing toll architecture. Since provincial border toll stations were removed in 2019, gantries have tracked vehicle paths across a connected national system.

A 2022 official tolling explanation described the chosen architecture clearly. It combines electronic nonstop collection, plate-image recognition, and multiple payment methods.

In other words, cameras are already part of ETC operations. The new plan changes which vehicles can rely on them as the primary credential.

That is an important shift, but it is not a clean break with the existing network. China is adding a camera-first service above infrastructure that already mixes radio and visual evidence.

The next question is why the ministry remains cautious when the necessary cameras and payment platforms already exist.

Why ETC Still Has an Advantage at Highway Speed

ETC separates vehicle identification from visual conditions, while a camera-first system makes the image the center of every transaction.

A standard ETC transaction starts when roadside equipment communicates with a vehicle’s onboard unit. The radio exchange identifies the account through a dedicated credential.

Cameras can then support the transaction. They can check the plate, classify the vehicle, preserve evidence, or help resolve an exception.

In a plate-first model, the order changes. The image becomes the main key that connects a physical car with its route and payment account.

That workflow requires several systems to agree. A camera must capture the plate, software must read every character, and a database must find the correct registration.

The system must then connect that registration with an authorized payer. Finally, it must match separate observations into one continuous trip.

Each step works today. The engineering problem comes from multiplying their failure rates across millions of vehicles and numerous toll points.

Cameras face ordinary road conditions. Dirt, snow, rain, glare, shadows, damaged plates, temporary obstructions, and poor mounting angles can reduce image quality.

Vehicle spacing creates another challenge. A camera must distinguish adjacent cars while connecting the correct plate with the correct axle count and lane position.

That relationship matters because toll calculation can depend on vehicle class and route. A correctly read plate attached to the wrong vehicle still creates a billing error.

An ETC transponder provides a machine-readable identity designed for this environment. It does not need software to infer characters from a changing image.

That does not make ETC perfect. A device can fail, lose power, become detached, or remain linked to outdated account information.

Radio reads can also require camera support. Toll operators keep visual records because no single sensor catches every event without exceptions.

The difference lies in the expected exception volume. A hybrid system can send uncertain cases into reconciliation without making every trip depend on image recognition.

A camera-first system must treat image quality as a core service requirement. It also needs a larger process for ambiguous reads, unregistered plates, and failed payments.

Published computer-vision research illustrates this gap between test performance and full deployment. One toll-focused study reported plate-reading performance above 98 percent on its selected dataset.

Other benchmark research has produced lower results under different datasets and assumptions. The numbers cannot be transferred directly into a national tolling guarantee.

Even a 98 percent result leaves two uncertain reads per 100 observations. A long trip can pass multiple gantries, creating repeated opportunities for errors or missing records.

Operators can combine observations and use confidence scores. They can also compare plate reads with vehicle color, make, model, axle count, and travel sequence.

Those tools improve reconciliation, but they add computing and review costs. They also introduce new ways to associate an observation with the wrong account.

China’s scale makes small percentages consequential. The Transport Ministry reported that the country had 199,400 kilometers of expressways at the end of 2025.

Its transport statistics also recorded 49.733 billion noncommercial passenger-car trips during 2025. That category grew by 4.7 percent.

Not every one of those trips used a tolled expressway. The figure still demonstrates the volume surrounding China’s road-payment infrastructure.

During a holiday surge, a minor delay can become a queue. A modest increase in disputed transactions can also become a major customer-service burden.

ETC minimizes roadside interaction by using a dedicated credential. That helps explain why the ministry still recommends it for frequent highway users.

The new phone-based service solves a different problem. It gives occasional users a lower-friction path without forcing them into an ETC enrollment process.

That path can succeed without immediately matching every ETC characteristic. A driver who avoids a physical card may accept stopping briefly to complete payment.

The early trials can measure whether that compromise works. They can also show how often camera reads require intervention under real traffic conditions.

Until those results exist, replacing ETC would exchange a known hardware burden for uncertain operational costs. The burden would move from the windshield into cameras, databases, support teams, and enforcement.

ETC Versus License-Plate Payment Is an Identity Tradeoff

The real contest is not radio versus cameras. It is a dedicated vehicle credential versus a publicly visible identifier.

A transponder exists specifically to participate in a tolling system. Its account relationship can be authenticated, updated, suspended, and checked against the vehicle.

A license plate serves many purposes beyond payment. It is designed for public vehicle identification, not secret account authentication.

Anyone near a road can see or photograph it. A dishonest driver can obscure a plate, alter a character, or display a copied registration.

Those risks do not prevent plate-based tolling. They change how much evidence, enforcement, and dispute handling the operator needs.

Consider a cloned plate. The camera may read every character perfectly, yet the transaction can still reach the wrong registered owner.

Computer vision cannot solve that problem alone. The image contains a valid-looking identifier, so the system needs secondary vehicle characteristics or human review.

A payment account introduces additional questions. The operator must confirm who authorized the plate connection and what happens when ownership changes.

Rental cars, corporate fleets, leased vehicles, borrowed cars, and recently sold vehicles complicate the account relationship. Each can separate the driver, payer, and registered owner.

A national service must define responsibility for every variation. Otherwise, convenience for one user becomes a billing dispute for another.

Payment failure is another difference. A transponder account can be checked against an established tolling relationship with known collection rules.

A plate linked to a digital wallet depends on current authorization, sufficient funds, and accurate account status. Revoked consent can interrupt the workflow after travel occurs.

The road operator cannot simply reverse the journey. It must pursue payment, assign liability, and preserve evidence for a possible challenge.

International systems show that license-plate tolling works best with substantial back-office support. Ontario’s 407 ETR offers both transponder and camera-based travel.

Its all-electronic system photographs vehicles without a detected transponder. The operator matches entry and exit images before billing the registered owner.

The road remains free-flowing, so the example proves that physical tollbooths are not technically necessary. It also reveals the economic difference between the two credentials.

Camera users incur separate administrative treatment, while transponder users avoid camera-related account handling. Frequent users are encouraged to retain the device.

Britain’s Dartford Crossing uses a related approach. Drivers provide a registration number and can connect it with an account and payment card.

The official Dart Charge service requires payment by a deadline when automatic account collection is unavailable. Late or unmatched payment can lead to a penalty process.

Neither system makes plate-only travel effortless behind the scenes. Both depend on vehicle registries, billing rules, image processing, notices, and dispute procedures.

China has an advantage because mobile payment is widely available. The planned service can place account setup and payment inside a familiar smartphone workflow.

However, payment convenience does not improve the original camera image. It also does not prove that the photographed car belongs to the account holder.

This is why plate recognition should be understood as evidence, not authentication by itself. The characters identify a claimed registration, while other systems establish responsibility.

ETC provides a stronger transaction credential because the roadside system reads a registered device. The plate image can then operate as a cross-check.

Removing the device shifts more trust toward databases and visual inference. It also increases the value of secondary attributes such as vehicle class and appearance.

The ministry’s staged rollout is therefore a rational identity test. MTC lanes provide a controlled environment where vehicles already slow or stop.

Slower movement improves image opportunities and allows drivers to resolve payment problems before leaving. Staffed or supervised facilities can handle early exceptions.

ETC-only lanes create a stricter challenge. They are designed around uninterrupted movement and rapid radio transactions.

Before plate-first users enter those lanes, the ministry needs evidence that the system can manage ambiguous identities without causing unsafe stops.

This opponent map explains the policy more clearly than a hardware comparison. The device is not surviving because cameras are primitive.

It is surviving because a purpose-built credential narrows uncertainty. Plate payment transfers that uncertainty into software, enforcement, and customer support.

What License-Plate Payment Risks Hiding

Removing a visible device does not remove infrastructure. It makes the remaining infrastructure less visible to drivers.

A camera-first toll network needs high-quality imaging at every relevant control point. It also needs lighting, lane sensors, connectivity, storage, and redundant computing.

The system must preserve enough data to support billing. It may also retain images for fraud detection, enforcement, and dispute resolution.

That creates a privacy issue beyond payment. A connected sequence of plate observations can reconstruct where a vehicle traveled and when.

China’s current ETC gantries already support route reconstruction for accurate charging. Expanding camera-first access raises the share of trips primarily tied together through images.

The policy question is not whether toll roads should observe vehicles. Toll collection already requires route and identity information.

The harder question concerns data boundaries. Operators must decide what they retain, how long they retain it, and who can access it.

Britain’s Dart Charge publishes a privacy policy covering plate images, crossing records, account data, and registered-keeper information. That documentation illustrates the governance required around camera tolling.

A comparable service needs clear rules for mistaken reads. Drivers should know how to challenge a trip that belongs to another vehicle.

The system also needs controls around account linking. A person should not be able to attach someone else’s plate to a payment account without adequate verification.

Vehicle sales create a predictable risk window. The old owner, new owner, registry, toll account, and wallet provider may update their records at different times.

Without synchronized changes, tolls can reach the wrong party. Automated collection can make that mistake feel immediate and opaque.

Network security matters as well. A transponder database is sensitive, but a plate-payment platform connects several broader systems.

It may connect vehicle records, route observations, telephone numbers, payment credentials, and identity-verification data. That combination has greater value if exposed.

System designers can reduce the risk through tokenization and separation. A toll operator does not need unrestricted access to a driver’s underlying payment details.

Payment providers can return authorization tokens rather than raw card information. Access logs can document which employee or system queried a travel record.

Retention limits can reduce the amount of historical movement data available. Encryption can protect records in transit and storage.

None of these controls is automatic. They require policy, implementation, audits, and credible consequences when systems misuse data.

Equity presents another issue. A phone-centered service assumes access to a supported device, a payment account, and enough digital confidence to complete setup.

The ministry is preserving MTC and ETC choices, which reduces that concern. Drivers can continue using established channels instead of joining the new service.

The rollout also targets small passenger vehicles first. That avoids immediately applying camera-first billing to every truck configuration and commercial fleet workflow.

Heavy vehicles can require more complex classification. Plate recognition alone may not reveal the axle or trailer configuration needed for a correct charge.

Even passenger vehicles produce edge cases. Temporary plates, diplomatic registrations, foreign vehicles, emergency vehicles, and damaged plates need separate rules.

Weather and maintenance can amplify these risks. A clean laboratory image does not represent a wet plate under glare at night.

The system can use several cameras and illumination angles. It can combine front and rear observations where regulations and plate formats allow.

Yet each additional sensor creates maintenance obligations. A degraded camera must be detected quickly before it produces a long batch of uncertain transactions.

The greatest risk is not a dramatic system failure. It is a steady stream of small mistakes that users struggle to understand or correct.

That stream can weaken trust even when the overall recognition rate looks impressive. Billing systems are judged one disputed transaction at a time.

This is also why a nationwide replacement would be premature. ETC and camera evidence can check each other during the transition.

If the radio credential and plate agree, the system gains confidence. If they conflict, the event can be flagged before automatic collection proceeds.

The “phone plus” trials should retain that cautious approach. They should publish operational measures instead of relying on broad statements about convenience.

Useful measures include successful plate reads, corrected classifications, payment failures, disputed charges, and average lane processing time.

The ministry has not publicly supplied those trial results yet. Its 2026 deadline concerns service coverage, not proof that plate payment can replace every ETC transaction.

That distinction should remain central. Availability shows that a service exists, while reliability shows whether it can carry the network’s full burden.

Three Signals Will Decide Whether ETC Devices Can Disappear

China’s trials must prove operational reliability, trustworthy dispute handling, and true nonstop travel before plate payment becomes an ETC replacement.

The first signal is the MTC rollout across the Chengdu-Chongqing region and the Yangtze River Delta. Coverage alone will provide only a partial answer.

The stronger evidence will come from transaction quality. The ministry and road operators need to measure how many eligible trips complete without a card or staff intervention.

They also need to distinguish several outcomes. A successful image capture is not the same as a correctly reconstructed route and completed payment.

If most eligible drivers complete the process accurately, the case for broader plate-based access will strengthen. High intervention rates would support ETC’s continuing role.

The second signal is the quality of exception handling. Every large toll system encounters unreadable plates, outdated registrations, failed payments, and disputed vehicle matches.

A mature service should let drivers inspect trips and correct errors through a clear process. It should also pause collection when evidence remains uncertain.

Fast dispute resolution would show that camera-first billing can scale beyond a demonstration. Persistent complaints would reveal costs hidden by the missing device.

The third signal is expansion into ETC-only lanes. This is the decisive test because those lanes embody the original promise of electronic tolling.

A vehicle should enter, travel, and exit without collecting a card or stopping for payment. The system should identify it reliably at normal operating speed.

If plate-first users can do that during peak traffic, the difference between the two services will narrow. ETC would then compete mainly on confidence, account simplicity, and operating cost.

If those users still need to stop, the new model remains a better MTC experience. It would not yet be a functional substitute for ETC.

International examples suggest that a mixed system can persist for years. Ontario’s camera option did not eliminate the transponder’s advantages for frequent travel.

Britain removed physical payment booths at Dartford, but it replaced them with account management, deadlines, and enforcement. The hardware burden moved elsewhere.

China may reach a different balance because its network, mobile-payment adoption, and state vehicle registries have different characteristics.

Its existing gantries also provide a significant foundation. Cameras already help identify routes, verify transactions, and investigate toll avoidance.

That makes gradual convergence plausible. ETC devices could become less necessary for occasional passenger-car travel as camera-first coverage improves.

A universal phaseout remains a much higher bar. The replacement would need to handle frequent commuters, fleets, heavy vehicles, exceptional registrations, and weak network conditions.

It would also need a transparent privacy framework. Travelers should know how plate images become payment records and how long those records remain available.

The ministry’s response leaves that future open. It does not frame ETC as permanent, but it treats the device as the current benchmark for nonstop travel.

That position avoids a false choice. Drivers who value maximum speed can retain ETC, while occasional users gain a lower-friction cardless option.

Competition between the two routes can produce useful evidence. Adoption rates will show whether installation is truly the main barrier.

Processing times will show whether cameras can preserve lane capacity. Dispute rates will reveal whether software convenience survives imperfect real-world data.

The most revealing comparison will concern total operating cost. A free or easily obtained device can still be cheaper than repeated imaging, reconciliation, and collection work.

Conversely, continued improvements in cameras and automated matching can reduce those costs. Integrated vehicle systems may also provide new credentials without a separate windshield unit.

For now, license-plate payment is feasible, but feasibility is not equivalence. A working payment path does not automatically match a national ETC network.

China’s 2026 rollout should be read as a controlled migration of selected drivers away from physical cards. It is not an order to remove every ETC device.

That narrower goal still matters. It can make infrequent highway travel easier while generating the evidence needed for a larger decision.

Drivers and transport operators should watch the trial results, the dispute process, and the first ETC-lane expansion. Together, those signals will show whether the device is becoming optional infrastructure.

Until then, the most accurate answer is conditional. License-plate payment can work, and China is preparing to use it more widely.

ETC remains harder to replace because it solves identity before the camera must infer it. The new system has to prove that software can close that gap at national scale.