Concerning the global initatives, more could be done sooner IMHO - but it's a start.
Concerning proposed plans from Dr Ashurebyli, I didn't really notice any. It's pretty difficult to comment on "satellite constellations" when there's no details concerning distances and function of constellations or the mechanism upon which this protection should operate.
The early detection of asteroids is certianly a problem, with only about 1800 objects being currently tracked with multiple "near misses" so far. This likely does not even represent 50% of the available threats in play currently - and more leave the Oort every day. As previously mentioned, the first time these are spotted it's commonly provided an incredibly short window of reaction time. Unfeasibly short. Also as previously mentioned, expanding sensor networks and better detection techniques is the obvious solution that can lead to greater reaction times being applicable. One would assume that this represents at least the first phase of these "satellite constellations". I seem to think we'd want birds just a little further out than the moon and once they're looking in more directions than they are not, starting to place some around sol at a similar orbital distance to Earth - A series of these in this belt should provide even more data - if that still doesn't provide enough sensor coverages then at a similar orbital belt to Mars should give much earlier detection times.
With regards to mitigation techniques then as previously mentioned, there's commonly not enough time to get a launch together, let alone line up a launch window. The obvious solution here is to have one ready to rock - and preferably already in orbit. If already in orbit the deployment time is drastically reduced, allowing for potentially earlier intercept. If you can get it there early enough then a nuke can be detonated before it hits. Rather than use the kinetic impact to reduce it's size, slow it by a few meter per second or nudge it to the left slightly - far enough out will mitigate a collision. If going for mass reduction multiple nukes would be required - after taking the larger rocks down in size(possibly requiring several nukes, one to split in three, three to hit them, 12 to hit that debris) setting a nuke off just in front of them could either vapourise them entirely or provide for more than enough deflective forces.
To assume there is enough time to attempt an earlier intercept (ie: most things we've already seen and are tracking) then to assume this can be done early enough we can begin to play with mitigation techniques. I agree heavily with "trial runs". For sake of the example, lets target 2017 DQ36, this has 8 opportunities to collide with Earth between 2021 and 2097, at about ½km radial it's a baby rock on the scale so perfect for playing with mitigation techniques, it's relatively low mass would allow for easier mitigations. With a Palmero scale of -3 then any failure in the attempt isn't likely to be overly hazardous. It's an ideal candidate.
An asteroid, even one as small as that, is unlikely to "stopped in it's tracks" as accurately pointed out previously - but nothing is static in space, and if you was to become, then everything else is still a hazzard as that is all moving. Luckily stopping it isn't required. All you need to do is slow it down(or speed it up, which can be easier depending where in it's orbital elipse it is) - when two moving objects have intercepting trajectories then adjusting the speed of one object should allow the other to clear the intercept zone so both these objects don't try to occupy the same space at the same time. Equally, if it's far enough away then adjusting it's heading by but 1° can over the length of it's arc provide for sufficient adjustments - but commonly you'll get the most overall effect for the minimal energy from adjusting velocity. And even more by doing it at the "right" time.
As for how to provide that course adjustment....
NASA have had some detailed studies on this. The tl:dr; of the matter is chemical explosives to obliterate are unlikley to be sufficient to make into small enough fragments, and would likely require embedding to stand a chance of fragmenting (but they was thinking a lot bigger than our example, we could feasibly blow something that small). Chemical explosives are unlikely to deliver sufficient kinetic energy to "displace". Nuclear explosives can deliver the yield, buried would provide a high fragmentation risk, as some large chunks are likely acellerated towards. The general consensus of the document was that it would be generally "better" and all around "easier" to detonate with some proximity in order to bounce more than break - almost like a gigantic game of pool, chipping it off into the corner pocket. Ofc asteroid composition and materials matter muchly. This sort of solution on a "gravel-type" asteroid is likely to make a serious mess and a lot of smaller problems.
For coating a single side with ice, and using sunlight to create "steam thrust" then it's unlikely to vent with any particular pressure, even over a wide surface area that low pressure is likely to be insufficient to provide suitable Delta-V and the largest effect would be adjusting it's center of gravity, which will cause(or more likely adjust) it's rotation. The same water in resevoirs - these could potentially be drilled into the face of the asteroid, or attached to surface - can then boil up and pressurise. Steam at pressure will give a lot more thrust from the same mass. Such systems are difficult to control. If the center of thrust doesn't balance center of mass then, depending how far off it is, it tends to push you in circles more than forwards and using Sol as thermal source you're limited in the directions you can apply thrust. Aiming is difficult, and requires a really early intercept.
A "salty chemical" is unlikely to compete with the cold of deep space - which is why the coolant loops on ISS use ammonia, it's harder to freeze that - Focus of a few lasers or microwave beams would be more reliable IMHO. Volumatic reduction is unlikely to be sufficient to be notable, unless the composition is mostly ice. Again applicable thrust is questionable, this is commonly delivered at pressure to provide speed. As is questionability to control how/where the thrust is applied.
Tethering chemical rockets and towing the asteroid could be considered viable to move something if it's hit far enough out, and you laid down thrust at the right point of it's orbital arc you could stand a chance of "steering" enough to avoid impacts. Several units may be required to stabilise rotation and prevent "yo-yo" or pendulem. I personally favour the likes of a net, for large solids and something resembling a wind sock for gravels, suspended between three or more units working in concert. Surface attachment would again require center of thrust to align center of mass to prevent spin. If far enough out then Ion propulsion systems can potentially provide a small impulse over a long period of time - and that tends to add up. If the EM-Drive is valid then this would be the most sensible type of application as it can continually output, and would possibly give rise to parking these problematic rocks for mining. Most "solutions" leave the problem there, in some form, they will be experienced again. At some point in time. Mitigation is obviously the primary goal, mining these to nothing is not an unworthy consideration as it would turn a problem into a solution.
"Bending it's trajectory sideways" is unlikely to be fuel efficient (Should get yourself a copy of KSP if you like this sort of thing, their orbital trajectory tool when you're plotting up manipulating orbits would demonstrate this concept visually, easily, and you can even play with various sized rocks that come close and building intercept systems. NASA chipped in some code especially for this). If you apply thrust just before the apoapsis - As far from center as it will reach - then the object is in theory travelling as slow as it will (naturally) and should require the least energy to result in adjustments, either way - but as it's just before it's "peak height" gravity is still slowing the object down as opposed to speeding it up so attempts to slow it are maximised. Conversly, just before the periapsis - The closest to center the orbit reaches - provides opportunity to have gravity assist in speeding up said object. If attempting to "go sideways" then at the periapsis/apoapsis provides the opportunity for the most effect(emphisis on the apoapsis). Attempts at other points is essentially wasting fuel, and without a propellantless system it's going to be difficult to achieve the required effect anywhere else, or take a lot of fuel to do so. Speeding up/slowing down gives you the most bang for your buck. When you're at the apoapsis then speeding up/slowing down should "push" the periapsis further away/closer to the body being orbited, whilst application at the periapsis should manipulate the distance of the apoapsis.
Solar sails are likely to be too low impulse to be viable for the larger percentage of examples, even with early intercept.