Modern trains generate far more data than their radio systems were first designed for. Onboard CCTV, passenger Wi-Fi, condition monitoring and passenger information all need a link between the moving train and the trackside network. Several technologies can provide it, and each makes different trade-offs on cost, ownership and performance.
What the link has to carry
- Live CCTV from cabins and driver cabs, for incident response
- Bulk offload of recorded video, at stations, termini or depots
- Passenger Wi-Fi
- Diagnostics and condition monitoring for maintenance
- Passenger information and announcements
- Backup connectivity supporting signalling and train control
The options compared
| Spectrum | Ownership and running cost | Behaviour on the move | |
|---|---|---|---|
| Standard Wi-Fi | Unlicensed | Operator owned; low running cost | Slow roaming and single-radio handoffs lead to disconnects and packet loss |
| Private LTE / LTE-R | Licensed | Operator owned; high capital cost and a complex rollout, with core infrastructure to run | Designed for mobility |
| Public 4G / 5G | Mobile operator’s | Shared network; recurring data charges and no network ownership | Inconsistent coverage, particularly in tunnels; shared bandwidth |
| Dual-radio wireless mesh | Unlicensed | Operator owned; no licence fees | Make-before-break handover, typically within 5 ms |
Standard Wi-Fi
Wi-Fi access points along the track are inexpensive and use unlicensed spectrum. The problem is roaming. A single-radio client has to disconnect from one access point before joining the next, and at train speed those gaps happen often. Controller-based roaming also adds a central dependency.
Private LTE and LTE-R
A private LTE network gives the operator a dedicated, mobility-ready system, but it needs licensed spectrum, base stations and a core network. That is a large investment for non-safety services such as CCTV and passenger Wi-Fi.
Public 4G and 5G
Using a mobile operator’s network avoids building infrastructure, but bandwidth is shared with the public, coverage is often weakest in tunnels, cuttings and remote corridors, and every gigabyte is charged. Uplink is the weak direction on public networks, and CCTV is almost all uplink.
Dual-radio wireless mesh
A mesh corridor uses unlicensed spectrum and operator-owned nodes along the track. Anywhere nodes on the train use two radios: one stays connected to the current trackside node while the other joins the next, so handover is make-before-break and typically within 5 ms. Each train can carry up to 700 Mbps at 80 MHz, and front and rear mobile nodes together can provide 1 Gbps or more. Nodes make their own routing decisions, so there is no central controller, and the trackside mesh heals itself if a node fails.
Combining technologies
Many operators use more than one system. Safety-critical signalling usually runs on its own dedicated, certified radio system, while a high-capacity mesh carries CCTV, passenger Wi-Fi and diagnostics and provides backup connectivity in support of signalling and train control. Choosing the right tool for each service keeps costs down and avoids loading the signalling network with video.
Onboard equipment
Whatever the technology, equipment on the train must meet rolling-stock standards. The AX52e holds EN 50121-3-2 for EMC, EN 50155 for railway electronics, EN 61373 for vibration and shock, and EN 45545-2 for fire protection. The same node is used at the trackside.
In service
Hong Kong’s MTR uses Anywhere multi-hop virtual fibre for railway infrastructure and wireless mesh for CCTV at stations. On the Danyang Monorail in South Korea, mesh mobility nodes on the cabins and along the route deliver live IP camera video from the moving cabins.
Questions to ask
- Which services need to run while the train is moving, and which can wait for a station or depot?
- How much capacity does each train need at peak?
- Is licensed spectrum available, and at what cost?
- What coverage is needed in tunnels and cuttings?
- Who will own and operate the network?