My notion is not outdated, it's just that what you are thinking of is nowhere near the levels used in UGVs like robotic dogs.
What you're proposing requires more than cheap 2-D single line Lidars, requiring vertical movement in the laser emitter/receiver to create 3D imagery, it's a far cry from the relatively simple ones used in electric Cars due to the different operational environment which only requires obstacle detection. There's differences between hobbyist grade and military grade also, something similar (even just as simple) will cost $500+ if expected to last in rugged environments. It's not just the Lidars that is expensive in a robot dog like UGV, you also need a lot of computing due to the amount of data and expensive actuators due to the complex movements it needs to perform.
Like it's not doing anything that a equivalent flying drone can't do for much cheaper, any damage to the sensor system from battle will basically cost more to replace than a suicide UAV anyway so a dog like UGV is extremely difficult to justify in a actual combat situation. Tracked UGVs are fine as fire support/logistics, but having a expensive, low endurance and low survivability armed UGV which is not as flexible as infantry men is questionable to say the least.
