I'm thinking they should bring back sensor ambiguity. The ability to instantly see everything there is to know about another race from a long distance thermal scan is just silly. I am thinking of three tiers of detection.
Tier 0 coulld be invisible to sensors, either because the emissions are so small they aren't looked at or there is no emission reaching the sensors.
Tier 1 could be a sensor blip, 'unknown:IR:600W/sr', which would yield no useful information besides a general (read ambiguous) bearing, distance, type and intensity. The 'marker' would then be placed within in the general area (IE: within 1 million kilometres) of the actual object and not updated at regular intervals, making it effectively a contact lost marker. Occasionally, there could appear 'ghosts' (false positive) and 'shadows' (false negative), which would add a bit of mystery to my current campaign (Earth and 100 Tracking Stations). Maybe even make it so a poorly trained or disciplined ship does not always check their scanners in a sub-pulse.
Tier 2 you'd be able to see the outline of the ship, tell what race built it and what make it is (if known). You'd be able to determine the exact energy emissions, in W/m2, which could allow you to guess future Tier 1 contacts with said ship. You could determine engine power, shield strength and pretty much everything else currently detected in Aurora.
Tier 3 would provide everything possible about the target, depending on the sensor. A thermal sensor would be able to check for weapons on the outer hull, if their powered and (when their powered) how powerful they are. An EM sensor could detect things like capacitors that are charged (and from that fire rate), sensor systems, antennae, communications (orders) and maybe even search for internal computers. This would have to be for military scanners only and for extremely close ranges, to balance the game. Maybe 0-500,000km for large sensor ships early in the game.