Wait--Again??

I was just browsing the news, and I couldn't believe this one: There's been yet another wrong-way crash on the Dallas North Tollway. Like most of them, it's on the older section of the road between downtown and LBJ, and, like virtually all of them, it involved a suspected drunk driver:

Isidoro Camacho of Dallas was driving a white Pontiac Sunfire and heading north in the southbound lanes for about a mile when he collided with a black Chevy Cobalt about 3 a.m., authorities said.

The Cobalt driver, Holly Harding, 23, of Lewisville saw the Pontiac and was able to take evasive action to avoid a true head-on collison, said Trooper Lonny Haschel, of the Department of Public Safety.

Thankfully, this one wasn't a fatality. And granted, the biggest problem in a drunk-driving crash is, well, that somebody was driving drunk. But what is it about this section of the Tollway that causes this to happen so much more frequently than anywhere else?

My theories? Perhaps it has to do with the fact that the "older" section of the Tollway has below-grade crossings, so the intersections with the offramps (since there are few true service roads in this section) look just like any other intersection; you can't necessarily see that the Tollway is below you. Does that cause people( (especially at night, especially while impaired) to automatically assume that they can turn left at any stoplight, and then do so? Sure, they have big DO NOT ENTER and/or WRONG WAY signs posted at the intersections, but do they need to be bigger? Transportation experts have said that the "cattle guard" method, where a wrong-way vehicle's tires would be punctured, don't work at freeway (or even offramp) speeds.

Is everyone missing something here? Is there a solution that nobody has thought of yet? Feel free to add your thoughts in the comments.