Q - How far away are approaching comets detected today? How long time before they would pass Earth orbit?
A - About 30AU's. That's the same distance as neptune.
Halley's comet, which is 70.56 km/s has been seen in 2003 at 28AU's away, near it's 35AU's aphelion, using three telescopes for multiple nights, Halley is 11km diameter and currently has apparent magnitude 28v.
Here is a graph of apparent magnitude of a 15 kilometer comet called 2015_ER61 over 3 years from 2014 to 2017, which was detected by PANSTARRS (Panoramic Survey Telescope and Rapid Response System). It passed 0.1AU's away from us. PANSTARRS also saw the 50m interstellar asteroid in November 2017 as it was travelling away.
Pan-STARRS currently consists of two 1.8 m Ritchey–Chrétien telescopes located at Haleakala in Hawaii.
Exoplanets were rare when they were detected with apparatus similar to PANSTARRS, and then they launched Kepler to find more, and they have as of 1 December 2017, detected 3,710 planets in 2,780 systems.
The same will happen in the future with comets, especially for interstellar comets, which is like having material from many light years away arrive in the solar system.
The difficult task is approaching a 25KM/s comet at landing speeds.
Most of ESA's GAIA photography work is thrown away! When it could be used to detect comets! It discards most of it's CMOS data, including all the 1000ds of comets that it would have detected had it been designed differently.
The bottleneck in GAIA is how much data it can transmit back to our planet.
Here is a suggestion of a future mission which can possibly multiply our knowledge of comets and asteroids by 2 orders of magnitude, the suggested mission is called FAiRLiGHT based on our experience of processing large amounts of CMOS data for vectors.
100W of solar panels can supply a 10TB SSD type HDD in space.
10TB of SSD can keep a month's worth of space telescope observations on disk.
50W is also used for a serial CPU processor.
The CMOS data is kept in a virtual 3D or Voxel array inside the 10TB memory, with simple memory indexing that represents Pixel X, Y , Time values. (it's cubic memory, representing space as a 2D plane with a time variant)
That's mathematical simplicity. You keep super simple indexing for all the photos in the momory, so that the processor requires nearly zero instructions in order to read, write and recall memory very fast.
A 50W processor can process about 10 million CMOS pixels every second searching for vector data on board the FAiRLiGHT space telescopes. To be precise it can process about 100 MFLOPs, floating point operations per second.
It's a comet/asteroid reconnaissance mission that is essentially keeping all the GAIA photos in large memory banks and running a processor on 10 million pixels per second searching for shooting-star type signals across multiple photos.
The clearest vectors/comets are then sent back to earth, which results in a thousand fold compression of the original CMOS files.
They will be able to use Bi-focal and Tri-focal versions of GAIA telescope, which then triangulate exact positions for comets.
The vogue has recently been about algorythms to study peaks in stallar brightness which describe exoplanets, and if material from other stars can be detected more efficiently like exoplanets were, for example by using satellite based bifocal vector processing engines, vector processing will perhaps become a well studied field and comets and interstellar meteorites will be seen frequently and with precision.