One Firmware Upgrade for the Installed Base: What Low-Altitude and Embodied AI Teach Fleet Operators

The most transferable idea at the show was not the 185-kilogram spherical robot. It was a single sentence from a private-network vendor: officers already carry a radio, so upgrade that terminal instead of adding another device.

Low-altitude security and embodied intelligence were the two fastest-rising tracks at Security China 2026. For the in-vehicle video industry, however, the most useful lesson came from a quiet product decision made by a private-network communications vendor.

The Most Transferable Idea Was Not a Robot

Hytera, which holds more than half of the installed base in public-security private-network communications, chose to add Remote ID detection to its existing radios through a firmware upgrade. Its business lead explained the reasoning to China Industry News: officers already have to carry a communications terminal when they respond to a call, and adding a separate detection device is too much to carry; upgrading the equipment already in their hands supplies the missing capability. Its low-altitude management platform connects into existing command and dispatch systems to close the loop from detection to handling.

That path avoids three obstacles at once — a new procurement budget, extra carrying load, and fresh training and certification. It is also a close match for the fleet situation. By the end of June 2026, in-service vehicles in the regulated categories had been retrofitted with intelligent video alarm devices under national policy. What is on the road now is a very large installed base of terminals that are already fitted, connected and certified, but whose compute and algorithms date from their delivery year.

Vehicles Became Mobile Ground Nodes

Other exhibits point the same way. FanShuang showed vehicle-mounted and suitcase-integrated counter-drone equipment, with its lead noting that detection hardware has converged to the point where the competitive focus is the combined effect and price of ever smaller, cheaper and better-performing devices. China Railway Signal and Communication embedded low-altitude security into system integration: front-end sensing first determines legality, then data returns to a command system showing flight situation, threat level and protection zones; its railway drone escort monitoring system uses a vehicle-machine coordination model to identify intrusions such as track vehicles, rockfalls and animals in advance, with related systems in routine operation in Xiong'an, Hohhot and Longjiang.

Shanghai Tejin presented a city-level TDOA low-altitude management scheme using a high-plus-low station layout — rooftop stations covering above-building airspace and pole-mounted stations covering below-building airspace — to reduce detection blind spots in complex urban environments. It also showed an X2D drone detection and positioning device and an AI multi-source low-altitude sensing engine; two mandatory national standards for counter-drone equipment are being drafted with the company among the main contributors. Hope-Wish exhibited ship and vehicle optoelectronic systems, and its HP-DMA35 multi-function camera won an industry innovation award.

The embodied intelligence zone supplied the visual drama: a 185-kilogram spherical robot from Logeng (Hangzhou) capable of 35 km/h, amphibious operation, four tonnes of impact resistance and net launch, already in daily patrol use with police in Zhejiang and Fujian; and a robotic arm with tactile sensing delicate enough to pick up a crisp. The show also released a top-ten list of high-value public-safety patrol and inspection scenarios plus a 2026 white paper on quadruped robots in public safety.

Make No Hardware Swap a Product Strategy

Low-altitude security is a genuinely new market, but Hytera entered it conservatively: no new device, just a new capability on a terminal the customer already owns. For fleets, the parallel is strong. Municipal transport authorities have already named high false alarm rates as a reason to upgrade and optimise on-board intelligent video terminals. Replacing hardware means high cost, vehicle downtime and re-certification. Algorithm distribution, model OTA, edge containerisation and a small external co-processor are far more realistic upgrade paths — and they convert a one-time hardware sale into a recurring capability business.

There is a second, longer-term shift. When drones, robots, roadside units and command platforms start calling each other, the vehicle stops being a monitored object at the edge of the network and becomes a mobile ground node with its own power, connectivity, location and viewpoint — and the ability to move. That is the most underrated structural advantage of in-vehicle equipment. Wang Nan's phrase about moving from eyes and ears to brain and limbs can be translated for vehicles as moving from recording what happened to participating in the response on site.

Risk Note

A show-floor demonstration is not delivery capability. Interface standards, responsibility boundaries and command authority for multi-device coordination are not yet unified — the two mandatory counter-drone standards are still being drafted. Fleet buyers evaluating vehicle-machine-cloud coordination should ask for a list of systems in routine operation and reproducible joint debugging tests, not a concept demo. Likewise, RaaS cost models vary enormously with local labour costs and should not be copied across markets.

What to Take Away

• Define an explicit capability-upgrade roadmap for installed terminals: remotely deliverable algorithm list, compute ceiling, external co-processor option and rollback mechanism.

• Sell the upgrade capability itself — subscription or per-fleet licensing — rather than treating it as free after-sales support.

• Evaluate whether the vehicle can export events as a mobile sensing node, but settle interface standards and command authority first.

• Prefer projects already in routine operation, such as the low-altitude management deployments in Xiong'an, Hohhot and Longjiang, over show-floor parameters.

Our Judgment

Embodied intelligence and low-altitude security will grow faster than traditional security over the next five years. But the most realistic opportunity for the in-vehicle surveillance industry is not to build robots. It is to capture the capability upgrade of the installed base, and to position the vehicle as a callable node in a heterogeneous safety network. The first is revenue that can be realised within two to three years; the second is a longer-term place to stand.

Sources: China Industry News reporting on the embodied intelligence and low-altitude zones, the organiser's 2026 innovation shortlist and scenario/white-paper releases, a Hope-Wish post-show release, and the Jiaoyunfa 2026 No. 2 retrofit requirements. Vendor performance claims and deployment descriptions are public statements, not independently verified results.