Technology has since evolved, and signalling with it. Through the 20th century, semaphores were gradually replaced by coloured lights, for reasons of simplified maintenance, reduced cost, and increased train capacity. Further accidents led to extra controls to combat hitherto unforeseen edge cases. 

Whilst the very first signalling principles were almost exclusively developed in the UK, as railway technology was independently developed around the world, each jurisdiction began to create its own operating and signalling principles. All had the same ultimate goal of ensuring the utmost safety, but each began to achieve this in different ways. 

The underlying architecture, which proved excellent at ensuring safety and allowing an acceptable level of control, remained unchanged: lineside lights inform drivers whether it is safe to continue at a given speed, or whether they need to brake, based upon some situational processing performed in a lineside interlocking installation. 

Growth in traffic meant that networks and junctions became many times more complex. 

Simple mechanical interlockings were gradually ruled out in favour of solutions which used daisy-chained electro-mechanical relays. These more complex interlockings necessitated an ever greater specialisation amongst the growing number of experts required for their lifecycle, so much so that they formed their own engineering institution in 1912. Later, in the 1970s-80s, solid state processors (just like in a computer) were deployed into railway control systems in a world first for safety engineering. 

The first colour light systems were revolutionary when deployed in 1898 on London Underground.  Despite the basic idea now being over 125 years old, colour light signals control the majority of signalled track mileage worldwide. Still, as the old adage goes, 'if it ain't broke, don't fix it': and colour light signalling is definitely not broken! So why are Universal Signalling proposing something different? We'll come to that later.

RETB (Radio Electronic Token Block) was designed as a lower-cost system for long single line sections with low traffic. It controls many parts of the Scottish network like the far north line. Though low-cost, its architecture limits both capacity and operational flexibility, and prevents wider rollout. The latest iteration is RETB NG (Next Generation), which is reportedly much more reliable than earlier implementations.

CBTC (Communications-Based Train Control) was primarily designed as a metro control system where the priorities are short headways between trains to give a high service frequency. Its main drawbacks are its expense, that it requires an integrated cab fitment, and is only really suitable for a captive fleet. 

ETCS (European Train Control System) is the specified UK mainline successor to all existing signalling systems. In the UK ETCS has been re-branded as 'digital signalling', mainly to make an easier soundbite for the media, but this is a misnomer: Signalling has been digital since the very first mechanical interlocking computers in the 1830s. Digital, in this sense, refers to the digital (i.e. transmitting data, not voice) radio link to the train cabs.

ETCS is currently (as of 2023) being rolled out on small parts of the UK network in first-in-class projects. These projects are slated to take decades. Initial deployments, which are shaking down the technology and installation methodologies, cost what the whole rail industry has acknowledged is a completely unsustainable amount.