Auto Crash-Avoidance Technology Found to Reduce Collisions – ABC News

Posted on by cathywu

video platformvideo managementvideo solutionsvideo player

Safety features that help drivers avoid front-end collisions by braking autonomously and headlights that help guide vehicles around corners are, indeed, reducing crashes, according to insurance claim analyses by the Highway Loss Data Institute, an affiliate of the Insurance Institute for Highway Safety.

via Auto Crash-Avoidance Technology Found to Reduce Collisions – ABC News.

But analyses of another safety feature were not so promising. Technology that warns drivers when it is unsafe to leave their lane is, in some cases, increasing the number of crashes.

Lane-departure warnings, which tell drivers when they are drifting out of their lane, were associated with increased collision and property-damage liability insurance claims as well as increased injuries to occupants of the insured vehicle.

engineering works!

Bitap algorithm

Posted on by cathywu

The bitap algorithm is an approximate or exact string matching algorithm that is one of the underlying algorithms of the Unix utility agrep (approximate grep). What differentiates it from other string matching algorithms is its use of bitwise operations, that make it run super fast.

The algorithm for exact string matching works by encoding the pattern to be matched in a bit matrix whose dimensions are the length of a long [31] by the length of the alphabet. So, let’s work with a small alphabet of size 3. A pattern such as “abcc” would be encoded as

1111111111111111111111111111110 # pattern/a
1111111111111111111111111111101 # pattern/b
1111111111111111111111111110011 # pattern/c

see more »

In the “a” row, there’s a 0 for every instance of “a” in the pattern, as indexed into the corresponding bit array. The same applies to “b” and “c.” The letter “c” occurs in the 3rd and 4th place of the pattern, so the 3rd and 4th lower order bits are set to 0. Everything else is a 1.

Let’s match this pattern to the text “aaabcca,” which requires a separate bit array for keeping track of how far we’ve gotten, if we need to start over, etc. More precisely, this bit array keeps track of all matches of the current iteration. We initialize this bit array to ~1 (not 1, long format)

1111111111111111111111111111110 # state

Now, we iterate through the characters of our text “aaabcca.” Here’s where the bitwise operations come in. We 1) select the first letter of the text and bitwise-or the corresponding pattern bit matrix with our state and 2) left shift the state over 1:

Text aaabcca
1111111111111111111111111111110 # pattern/a
1111111111111111111111111111110 # old state
1111111111111111111111111111100 # new state

Awesome, a 0 indicates a match and a 1 indicates a mis-match at the corresponding position. So, so far, we’ve got a match of “a” and are ready to see if the next letter matches too. Next:

Text aaabcca
1111111111111111111111111111110 # pattern/a
1111111111111111111111111111100 # old state
1111111111111111111111111111100 # new state

We’ve still just got a match at “a”…

Text aaabcca
1111111111111111111111111111110 # pattern/a
1111111111111111111111111111100 # old state
1111111111111111111111111111100 # new state

OK, I’m getting bored…

Text aaabcca
1111111111111111111111111111101 # pattern/b
1111111111111111111111111111100 # old state
1111111111111111111111111111010 # new state

Text aaabcca
1111111111111111111111111110011 # pattern/c
1111111111111111111111111111010 # old state
1111111111111111111111111110110 # new state

Text aaabcca
1111111111111111111111111110011 # pattern/c
1111111111111111111111111110110 # old state
1111111111111111111111111101110 # new state

Suppose the pattern is of length m. Each iteration checks the new state for whether the m+1th lowest bit is a 0. If it is, that means the pattern was matched to completion, like in this case! Awesome, we win.

Due to the limitation of this algorithm to strings of 31 characters or less, the runtime is O(1). The number of bitwise operations required is 2n, where n is the length of the text.

Sharing data links in networks of cars – MIT News Office

Posted on by cathywu

A new algorithm lets networks of Wi-Fi-connected cars, whose layout is constantly changing, share a few expensive links to the Internet.

via Sharing data links in networks of cars – MIT News Office.

“If you were trying to analyze vehicle traffic, there’s probably 10,000 cars on the Los Angeles Freeway that know that there’s a traffic jam. You don’t need every one of them to tell you that,” he says.

Alejandro Cornejo, MIT

For loops

Posted on by cathywu

The following C code iterates over the characters of a string. I’ve never thought to write code this way, but of course you can.

#include <stdio.h>

int main() {
    const char *text = "testtest";
    int n = 0;
    int i;
    for (i=0; text[i] != '\0'; ++i) {
        n++;
    }
    printf("%d\n", n);
}

Source: Wikipedia

Honeywell Wants to Teach Airliners How to Taxi – IEEE Spectrum

Posted on by cathywu

Honeywell is experimenting with an autopilot not to fly the plane but to taxi it, the Wall Street Journal reports.  It’s harder than it may seem because the biggest problem—sensing and avoiding objects—is particularly thorny in crowded airports.

via Honeywell Wants to Teach Airliners How to Taxi – IEEE Spectrum.

And that’s even harder than automating flight.

Interface for partially automating huge and bulky limited movement ground vehicles

Eventual applications to commercial airports, trucking industry, aircraft manufacturing, etc.

MIT Humans and Automation Lab (Prof Mary Cummings) [src]

AlaskaTrip

Posted on by cathywu

My adventuring up in Alaska has come to a close, and it’s time to return to the lower forty-eight. What started as an idle thought about a year ago has grown into a summer of driving companions for Tevis, snacking companions for Marie, Workaholics Anonymous therapy for me, 3600 miles on the road, and new experiences for all. In short, we drove up from the west coast through Canada to Fairbanks, drove up to the Arctic Ocean and the Prudhoe Bay oil fields and back, bused and hiked around Denali, and ended up here in Anchorage. I joined in for just an incredibly short but glorious 2 weeks, and they are continuing onward through more of Alaska/Canada and eventually back down to the Bay Area. Here are the highlights.

First, the animals

The final total:
25 caribou
14 bears (7 black, 7 grizzly, 0 polar)
8 dall sheep
6 mooses
6 foxes (some arctic, some red)
6 horses (wild)
And just about a million mosquitoes. Sadly, I lost count of how many times I got bitten.

Day 0

Tevis and Marie arrive in Seattle, home to the REI heaven! It is “huge and awesome” and meets Tevis’s expectations, who I take to be the expert on this kind of thing. There is a freaking waterfall and mountain bike trail on the outside of the storefront. In the center of downtown Seattle.

Day 1-4

Rain, rain, go away, come again another day. We drive and car camp through Canada and its miserable weather. On the plus side, I’ve now been to Canada! At one of the turnouts, we learn of this Chinese dude James Quong, a civil engineer who designed some 133 bridges (of ~200) on the Alaska Highway!

Day 5

We make it to Fairbanks, AK! The far away land of Alaska never seemed quite like a real place until now. That night, I fall in love with the Alaskan Amber.

Day 6-8

We drive up the 414-mile stretch of road between Fairbanks and Deadhorse, which has as far north as you can get by road in North America. The Dalton highway is a mix of paved, gravel, and dirt roads with vehicles of all shapes, sizes, and speeds going through. It’s the real deal with heavily limited service, and along the way, we attempt to learn trucker slang in hopes of understanding their conversations on the CB radio.

Marie succeeds in eating cheesy snacks through the swarms of mosquitoes gathered at the Arctic Circle sign post. I am delighted that we reach that point, but am simultaneously utterly disappointed that there is no painted white dotted line on the land marking the circle.

The mosquitoes are bad throughout the trip, but along the Dalton, they are truly truly numerous. There must have been hundreds swarming us as we set up camp and made dinner. I develop a mosquito dance in hopes of appeasing them, but it is only somewhat effective. I wish I could have captured a shot of the swarm because it was truly hilarious. We quickly give up on camping and cooking and opt to pile in the back of the SUV for the other nights.

In the land of the midnight sun, the sun simply does not set, and so, our sleep schedules went. There is no rush to get somewhere and set up camp before dark. Having the sun still up at 4am when we prepare to sleep is simply surreal. There is no other word for it.

We go on a tour of the Prudhoe Bay oil fields, which takes us through the area operated by BP. It’s strange to see all this heavy equipment next to freshwater ponds next to rows beyond rows of chemical packs next to caribou roaming the tundra. I’ve never seen a place quite like it; perhaps the camps out west seeking gold were once similar. We learn a bit about the engineering blunders of the drilling process and the 5K workers who work there. They liked to point out how environmentally friendly they are and all the things they are doing to prevent the permafrost from melting, but I’m not so sure about it all. The tour terminates at the Arctic Ocean where, with some persuasion from Marie, we all jump in and immediately regret it. Anyway, that’s 3 oceans touched, 1 to go.

Day 9

We head to Denali National Park in hopes of catching a glimpse of Mt McKinley through the weather and clouds it creates to hide itself from the world. The campground just outside Denali we stay at has internet. What a terrible thing for them to do to me.

Day 10-13

I finally get my hands on some reindeer sausage for breakfast at Denali. It’s subtly different, and I like it.

After sitting around eating junk food since the start of the trip, it is finally time to move. We set out on a 3-day, 2-night backpacking excursion up Mt Busia, just north of Mt McKinley. Soon after we summit, however, it proceeds to rain for a continuous 36 hours, so instead of day hiking the next day, we basically sleep and read for just as long. Just when we start dispairing at the miserable weather, the rain lifts around midnight on the second night, as did the clouds. Around 3:30am on the third day, the sun is up and there it is: the most perfect view of Mt McKinley. Absolutely no clouds obstructing the view.

We have seen what we came to see and are pretty fed up with rain, so we hike down a few hours earlier than planned. Due to the hours of rain, the river crossing we had done on the way in has become notably harder and more terrifying. Forming a line and going one step at a time, we cross the now-waist-deep and quick-moving 40-degree Moose Creek. Without the extra weight of my pack and Marie pushing me upstream, I’m sure the stream would have washed me away.

Day 14-15

The last stop is Anchorage, home to the Ted Stevens Anchorage International Airport, whose namesake uttered that the Internet is a “series of tubes.”

I’ve given you some words and pictures, but nothing is quite like going there and experiencing Alaska. Everyone we encountered on this trip, particularly the park rangers, was kind and helpful beyond expectation.