There is a kind of game that I've wanted to run, or play in, for a long time and the sad part about it is while most games have a way of dealing with the obstacles this game is based on, they don't have a way of doing it in the way they do for their primary conflict resolution (i.e. combat). What this means is that things that should be just as exciting as combat, often aren't due to the fact that games don't have ways of handling them. I don't have specific answers for these aside from "do your best, and be more descriptive" but I still wanted to talk to them a bit more. There are some specific suggestions though for each one, even if they're not necessarily answers.
By driving it could be any type of piloting really, but read a book or watch a movie with a good chase, driving, or barrelling down a canyon sequence and tell me that that isn't as exciting as the good fight scenes are. Now look at it in a Table Top RPG, you don't have the things for going back and forth like you do in combat, you can't deal damage and enjoy a critical. At best you can have a "give me 4 drive rolls, and pray you beat an X on all 4". That doesn't really get across the same level of tension these sequences are supposed to have in them though.
Some games, like Spycraft, have partial solutions for this with well thought out Chase rules, but what about other things? What if I want to fly my A-Wing down Beggar's canyon? That is supposed to be as exciting as a Death Star run, especially when you consider how often Tattooine farm boys claim dangerous canyon runs will be "just like beggars canyon back home". Clearly this isn't an event that should just be a single die roll.
So what should you do? Honestly, combat doesn't necessarily apply here for a mechanic but maybe we can borrow some stuff from it, ease up on the generic binary driving sequence of "everything is fine" and "oops, I wrecked my car". Go into it, describe the canyon, give the player options for how he can take things "I cut it close to shave time" or "I take it wide for safety", have those options impact the difficulty of the roll. If they make it, they make it (describe that too!) if they miss it don't just go into crashing but give their craft some damage (like if they took a hit in combat, don't drop them in one go whittle them down to size). Scale the damage by how much they failed the roll, and then give them an easier recovery roll to not fully crash. The journey then continues with them having banged up their ship, maybe even causing some system/part damage to it before they continue on.
This is a more generic 'player versus environment' one, the players are trying to scale a cliff or other obstacle. The main point of the game is the exploration and traversing across the weird terrain they've found. This may come up more often with more physically dangerous ways of traversal, such as climbing (hence the topic). The fun part about climbing though is that this can be turned into combat fairly easily.
First, look at the climb itself. Parts of it will be harder than others (higher difficulties to hit), and the mountain/tower/whatever totally has a numerical value that when over-come it is defeated in its height. So, take the height and assign it like hit points for the object. In return, the object and gravity are trying to kill the player right back, they're just more passive about it. Take these, and combine it with the option based thing from Driving and you have the chance for fun.
The player makes an option and rolls their climbing check, if they succeed they get to roll "damage" to see how much of the mountain they can scale with that check. Then it is the mountain's turn, make a roll against the player (use their climbing skill as their defense modifier for however your system figures it out) if the mountain hits, well then add a surprise on the next check for him. Make it harder (don't tell them), a loose rock, a slippery hold, something falling from above. Make it more difficult for them to succeed.
This makes the traversal actual player vs. environment, with the environment actually able to hit back. It also brings more involvement into a method of obstacle management.
Traps and Puzzles
Traps and Puzzles, as easy as they are to mechanically just solve (if the GM allows it) can be as frustrating as they can be fun. This is less a way to do this kind of thing, and more something you need to be cautious of. The trick with Traps and puzzles is how easy it is to figure out, you don't want it to be too hard but you also don't want it to be too easy. Where does the problem come in? In the clues you give out.
As the GM you know the solution, so clues that you leave will point directly to that solution because you know how it works. It's called inductive reasoning I believe, but since you are working from the solution to the problem, things will match up nicer for you.
On the other hand, your players just have clues. They don't see the end point for them, and with that lack they can get incredibly lost. So don't be stingy, dole out some clues and give them some roles. If you've given out your planned clues, and they're still lost 5 minutes later, give them another one. It gets frustrating working on a problem you can't find the answer to, and when you can't actually see the problem and are reliant on some words it can be even worse. So give some clues, guide them to the path you can see clearly, and if they have a clever idea that you never thought of, maybe give it a shot at working.
I just want to point out that these ideas are things for when your main obstacle for your game session or campaign is non-combat stuff. If your game focuses on combat, and the person just wants to climb a mountain or fly through beggars canyon to get to the battle, then just let them throw some dice and be done. If however, the player is proving they're good enough to get into a swoop gang by flying beggar's canyon, then build it up a little. Let them feel the tension as they race through the canyon. Always remember games are about having fun, and some of that fun comes from tension, but don't let that tension develop into frustration.