(Bloomberg) – A recent crash involving an UberTechnologies Inc. driverless car suggests autonomous softwaresometimes takes the same risks as the humans it may oneday replace. 

The accident on Friday in Tempe, Arizona, caused no majorinjuries. Another human-driven car turning left failed to yield,hit the Uber car and flipped it on its side. After a short pause,the company's self-driving test fleet was back on public roadsin Tempe, Pittsburgh and San Francisco early thisweek. 

Complex story

But the Tempe Police Department report, released Wednesday,recounts a complex story.

The Uber Volvo SUV, outfitted with autonomous driving sensors,was heading south on a wide boulevard with a 40 miles-per-hourspeed limit. It had two of the company's test drivers infront and no paying passengers. The light turned yellow as thevehicle entered an intersection. A green Honda on theother side of the road was trying to make a left at thelight. The driver thought it was clear and turned into the oncomingUber SUV, according to the police report.  

In a statement to police, Patrick Murphy, an Uber employeein the car, said the Volvo SUV was traveling 38 miles perhour, a notch below the speed limit. He said the traffic signalturned yellow as the Uber vehicle entered the intersection. He thensaw the Honda turning left, but "there was no time to react asthere was a blind spot" created by traffic. The Honda hitUber's car, pushing it into a traffic pole and causing it to turnon its side. 

Vehicle in autonomous mode

During the event, the Uber vehicle was in autonomous mode, aspokeswoman for the company and the Tempe policesaid. 

Others involved in the accident, though, didn't imagine a robotbehind the wheel. Alexandra Cole, the driver of the Honda, toldpolice that she could not see any cars coming when she decided tomake the left turn. "Right as I got to the middle lane about tocross," she wrote, "I saw a car flying through theintersection."

Another witness told police that Cole was not at fault. "It wasthe other driver's fault for trying to beat the light and hittingthe gas so hard," Brayan Torres told police in a statement. "Theother person just wanted to beat the light and keptgoing." 

Related: Will self-driving cars bring saferroads?

Eyewitness accounts can often be unreliable, and other witnessesin the police report did not say that the Uber car was at fault— something the police agreed with. Still, Torres'saccount raises the question of whether Uber's self-driving sensorsspotted the light turning yellow and, if so, whether it decided itcould safely continue through the intersection. One ofUber's self-driving SUVs ran a red light in San Francisco lastyear, and on five other occasions the company's mapping system forits cars failed to recognize traffic lights in the area, the NewYork Times reported in February. 

Potential hurdles

Uber's problems show the potential hurdles to winning approval for autonomous vehicles from thepublic and regulators. The company, and rivals like AlphabetInc.'s Waymo and major automakers, are working to tweak software tohandle "edge cases," like unusual driving conditions.

Self-driving cars have more often been criticized for drivingtoo cautiously, slowing or stopping when human drivers would bemore aggressive. Autonomous vehicles operated by Waymo have beenrear-ended due to such issues and the company has been working tomake its system more human. 

Related: Driverless cars closer to U.S. greenlight

There is a potential upside for Uber from the Tempe crash: Itnow has rich, unique data to use for its self-drivingprogram. Last year, after a Waymo car car bumpedinto a bus, the company said it used the incident, and"thousands of variations on it," to refine its software. "This is aclassic example of the negotiation that's a normal part of driving— we're all trying to predict each other's movements," itadded.