Not long ago, Xiaomi YU7 GT conducted an unmanned test at the New North Circuit, and from the results, the running was not slow. So does this mean that Xiaomi is becoming more and more confident in her intellectual driving? In this episode, we found a Xiaomi YU7 to see if its intellectual driving level has improved recently. The experience route is still our classic evening peak Qingbei line. The whole journey is 50 kilometers, 20 difficulties, narrow roads, congestion, overpasses, tidal lanes, high-speed toll stations, rich scenes, comprehensive investigation, even if it is driven by people, it is very difficult. So, can Xiaomi YU7, who came back from "taking the classics" in New North, make a convincing performance on the "Qingbei" route of our wisdom test evening peak? A sentence is a little unclear, let's leave a suspense first and say it at the end. Next, let's take a look at the specific performance of Xiaomi in 20 scenarios. Friends who would like a detailed video description are also welcome to click on the link below for the full details of the experience. More exciting videos can be found in AutoHome Video Platform Scene 1 | Self-contained basement Xiaomi YU7's had support parking space to parking space global navigation assistance, covering the park, basement and other internal roads. So after we learned the route in advance, it can start the pilot in the parking space, and then drive out of the basement independently according to the memory route. It drove pretty fast in the basement. If you encounter a right angle corner, you can borrow the opposite lane in advance and expand the turning radius to facilitate passage. On the uphill turn, Xiaomi YU7 encountered the opposite car and thought for a while before realizing that it was necessary to avoid detouring and judge correctly. Of course, there is room for improvement in route selection and logical judgment. When it came to the narrow position in front of the gate, Xiaomi YU7 controlled the route accurately and stopped smoothly at the toll window. After waiting for the pole to be lifted, the system also reacted quickly enough to start and leave on its own, and there was no manual takeover of the entire outbound warehouse. Scenario 2 | Unprotected left turn Unprotected left turn is a typical intersection game scenario. Vehicles should neither rashly speed up rushing, nor dry in place for a long time. It is necessary to accurately judge the emptiness of straight traffic flow and find a balance between safety and efficiency. In this scene, Xiaomi YU7 is not the first car at the intersection. The green light lights up and follows the car in front into the intersection. First, slowly move forward and wait for the opportunity. After finding the right vacancy, complete the left turn action. Performed well and handled naturally. Scenario 3 | Auxiliary road passage The auxiliary road passage scene seems ordinary, but implies three difficulties. First, the auxiliary road entrance is located at the door of the hospital, and there are often temporary stops for vehicles on the roadside. You can observe the performance of vehicles in identifying the auxiliary road entrance, dealing with occlusion, and avoiding illegal pressure diversion lines. Second, if the vehicle does not enter the first entrance, there is also an ramp 50 meters ahead, which can understand the selection judgment logic of the intelligent assisted driving system. Third, the signal light in the middle of the auxiliary road is very close to the stop line, which is easy to form a visual blind spot. Many models have made identification errors here. In this scenario, Xiaomi YU7 chooses to enter the auxiliary road from the ramp in front, but the speed in front of the ramp is still high, and the process of entering the auxiliary road is very fast. Although it successfully entered the auxiliary road, it relatively did not give the car occupants time to observe, which gave people some lack of confidence. Although the location of the subsequent auxiliary signal lights is relatively hidden, the Xiaomi YU7 can be stably identified and passed through the intersection. Scenario 4 | The pilotage scenario of the park pilotage park mainly depends on the fact that the vehicle support does not support the autonomous passage of ground roads in the closed park. The preset route requires vehicles to pass through the park's internal roads, and the entrance is easily blocked by electric vehicles, bicycles, and temporary parking vehicles, and the recognition is low. At the same time, there are gates at the entrance and exit of the park, which also tests the vehicle's ability to identify the timing of the pole lifting. Xiaomi YU7's had pilot assisted driving support park and basement have the foundation to complete this scene. In this scene, not only is the left turn somewhat concealed, but there are also vehicles that are parking on the side and turning around. Xiaomi YU7 first waited at low speed for a while, and then boldly turned in. Slow down in front of the inlet gate and accelerate through immediately after lifting the lever. The pilot assisted driving function in the park has not been affected in the slightest. Pedestrians facing the roadside will lightly dodge in the direction. Driving to the exit gate position is also recognized accurately. Slow down and wait for the lifting pole to pass. The Xiaomi YU7 performed well in the use of the park route and gate recognition, and the overall flow was very smooth. Scenario 5 | Left turn waiting area After entering 2026, identifying the left turn waiting area is already a basic function of the intelligent assisted driving system. In our preset first position, Xiaomi YU7 did not encounter the need to enter the left turn waiting area, but followed the normal left turn of the front car through the intersection, but it can be observed from the central control SR interface that the system has marked the left turn waiting area. In the next turnaround area, although the vehicle is still not the first train, there is still space in the turnaround area, and Xiaomi YU7 follows the front car across the stop line, indicating that the vehicle can understand the rules of the left turn turn turnaround area. Scenario 6 | Multi-select left turn In the multi-select left turn scene, the front fork road has two left turn passage methods: under the left bridge is a dedicated left turn lane, and the leftmost lane in the straight direction can also be turned left. Such a road design can improve the left turn passage efficiency of the single-wheel green light, while also observing the vehicle's ability to understand the navigation route and lane selection. The Xiaomi YU7 faced the fork in the road very decisively and drove directly into the underpass on the left side. However, under the bridge, there are few vehicles in the left lane, and the system is hesitant to change lanes. It has been tested many times, which has a slight impact on ride comfort. Then, after waiting in line for the green light, follow the car in front of you to make a smooth left turn. Scenario 7 | The residential area passes through the complex road section of the residential area, which is a common two-way single-lane in the urban area. There are two-wheeled vehicles, retrograde electric vehicles, temporarily parked online vehicles, and lateral road access vehicles on the road. The system needs to judge the movement of surrounding vehicles and pedestrians in real time, and dynamically plan the driving route. There are also risks such as "ghost probes" and quick insertion of takeaway vehicles in this section, and manual driving is very labor-intensive. In this section, facing the traffic participants on all sides, Xiaomi YU7 performed more calmly, avoiding nature on both sides of non-motor vehicles, and there was no excessive waiting. Overall, Xiaomi YU7 performed well in this scenario of passing through residential areas. Scenarios 8 and 9 | Turn right and merge into the main road + continuous lane change + under-bridge U-turn This is a combined scenario. Vehicles need to first observe the left-side motor vehicles and non-motor vehicles, and turn right at the appropriate time. After turning in, you must also continuously change lanes to the leftmost side, and complete the under-bridge U-turn. The three operations are not difficult to perform individually, but they need to be completed in a coherent manner over a short distance, which requires decision-making, coherence and lane change efficiency of the assisted driving system. Xiaomi YU7 slowed down in advance at the intersection and had a strong sense of defense. Then, while observing and probing, he found the opportunity to turn right and enter the middle lane directly, and then changed lanes again to prepare for a turn. At the turnaround position under the bridge, Xiaomi YU7 first expanded the turning radius to the right, and then entered the turnaround lane. Overall, the Xiaomi YU7 can complete the continuous actions of right-turn incoming, continuous lane change and under-bridge turnover, and the connection is smooth and the traffic efficiency is high. Scenario 10 | Non-hybrid reeling machine The non-hybrid reeling scenario mainly experiences the ability of vehicles to play with non-motor vehicles when entering and exiting the ramp in the reeling section. We manually adjust the waypoint to allow the vehicle to completely bypass the entire bridge. The Xiaomi YU7 was in the right turn position. Due to the large number of non-motorized vehicles in the rear, the system performed conservatively. It waited until there was almost no car before turning right into the ramp. During the subsequent bridging process, Xiaomi YU7 performed relatively smoothly. At the right turn position in front of the bridge, due to the occlusion of the right rear flower bed, there is a blind spot, Xiaomi YU7 has shown strong defensiveness, and it is still relatively conservative for non-motorized vehicles. At the time of the final transfer into the main road, Xiaomi YU7 was about to turn right into the second lane on the right, and suddenly returned to the rightmost lane, which had a slight impact on comfort. The Xiaomi YU7 will be relatively conservative in the face of non-motor vehicles in the non-hybrid bridge scenario, resulting in a slightly longer waiting time and some room for efficiency improvement. Scenario 11 | Riverside Path Riverside Path is crowded with traffic participants, and there are often temporary stops on the roadside. The core is to observe the ability of vehicles to handle when meeting. There are also two blind corners in the second half of the road. If the speed is not properly controlled and the observation is not timely, it is extremely prone to offset collisions. The Xiaomi YU7 is faster and travels smoothly in a relatively spacious position on this section. After entering the narrow road, the vehicle speed is reduced. In the face of the doorway of the community and blind bends, the system will have defensive actions, fully slowing down and passing through, so that the people inside the car can be more at ease. The slight flaw is that the braking action will be somewhat stiff, which will have a certain impact on comfort. Scenario 12 | Tidal lane The tidal lane will adjust the direction of traffic according to the change of traffic flow, which is used to relieve peak traffic pressure. At present, most models of assisted driving systems will not rush into the tidal lane, but there are not many models that can truly fully identify the status of the tidal lane, understand it, remind it, and use it reasonably. The Xiaomi YU7 has no special logo display in the SR interface, and the vehicle has not driven into the tidal lane on its own. During the process, Xiaomi YU7 once tried to enter the tidal lane, but due to the rapid approach of the vehicle on the left rear side, it did not enter. At the end of the tidal lane, Xiaomi YU7 was able to successfully select the correct lane. After turning around, the vehicle also independently avoided the current limited bus lane, and the overall performance was better. Scenario 13 | Three o'clock turnaround The three o'clock turnaround mainly observes the reversing control ability of the assisted driving system. In order to reduce the difficulty, we set the test site at the entrance of Wanliu Academy, which has plenty of room for borrowing. Nowadays, the three o'clock turnaround is a routine function of mainstream pilot-assisted driving. Setting this experience scenario is to screen systems that do not support this function temporarily. I hope that manufacturers will push OTA upgrades as soon as possible. The Xiaomi YU7 will voice the prompt "Please turn around manually" before the turnaround position we set, and then the system will be degraded. There are voice and voice prompts at the same time, so the current version of the vehicle does not have the three-point turnaround function. Scenario 14 | Very narrow road The very narrow road scene is a narrow road adjacent to the main road. When the vehicle enters, it needs to turn two right angles continuously. The road surface is very narrow. The right side of the road is still full of vehicles, and it may encounter non-motorized vehicles traveling in the opposite direction, and the traffic conditions are more stringent. The Xiaomi YU7 passes smoothly at the first right angle bend, while the second right angle bend cannot pass by itself due to the parking of the vehicle on the right, the space is relatively small, and the vehicle stops at the entrance. Therefore, the driver gently taps on the accelerator pedal, and at this time, the vehicle will still automatically control the direction to turn into the narrow road. In the narrow road, the millet YU7 will control the speed lower. Overall, in this scenario, the millet YU7 still has room for improvement. Scenario 15 | Inbound and outbound ramp Inbound and outbound ramp Inbound and outbound scenarios need to complete the switch between the Fourth Ring Main Road and the G6 Beijing-Tibet Expressway, including one ramp inbound and one ramp outbound. Among them, the exit is a three-fork intersection. Many models will enter the right-most ramp by mistake and miss the opportunity to get on the bridge. Xiaomi YU7 has been driving in the rightmost lane in advance, and the route to enter the ramp is also very steady. The driving trajectory is smooth. Facing the three fork intersection, the vehicle accurately drives into the middle ramp and passes smoothly. In the Panqiao section, the car is congested, and the Xiaomi YU7 is more comfortable to follow. When you get off the bridge, you will face the merger, and the car will perform smoothly. When preparing to exit the high-speed ramp, the main road traffic is large. After driving through the diversion line, the Xiaomi YU7 begins to test and then successfully remit. Scenario 16 | Self-passage of toll stations In order to achieve smooth assisted driving in all scenarios, in addition to the basement function, high-speed toll stations are also a breakpoint that must be overcome. Xiaomi YU7 is able to select etc lane passage or manual lane passage in the settings. Since the new car is not yet etc, we chose the artificial lane. Xiaomi YU7 chose the rightmost lane before approaching the toll station and drove to the toll plaza location. In the SR interface, the "artificial lane" will be prompted. At the same time, the vehicle is also correctly selected to the artificial lane. Although there is a prompt to take over, the vehicle can still pass by itself and accurately stop at the window position. Subsequently, when driving out of the high-speed toll station, the Xiaomi YU7 can also stop at the toll window on its own and accurately, in the right position, instead of driving all the way to the gate. After waiting for the pole to be lifted, the system prompts to gently step on the accelerator pedal to confirm the start, and after confirmation, the pilot assisted driving will continue to be connected, which is very good. Scenario 17 | The scenario of the roundabout under the overpass is very common in Beijing. The machine is non-mixed, there are many blind spots, and the passage pressure is large. Therefore, the gaming ability, route selection ability, and passage efficiency of vehicles in the roundabout are observed. The traffic flow in this ring island scene is relatively dense, and there are a large number of non-motorized vehicles crossing the intersection, and the road conditions are complex. Before entering the roundabout, Xiaomi YU7 slowed down and observed, and then seized the opportunity to enter the roundabout. It performed very smoothly in the roundabout and was able to choose a more efficient route, and the game was more natural. Overall very smooth and efficient. Scenarios 18 and 19 | Gaming out of the main road + ramp sharp curve Gaming out of the main road and ramp sharp curve is a coherent scenario. Due to the close distance between the entrance and the exit, the vehicle game needs to drive into the ramp, and the end is a three-way intersection. The two exits are side by side. Many models will accidentally drive into the rightmost ramp, missing the opportunity to get on the bridge. Xiaomi YU7 has been driving in the rightmost lane before approaching the exit. It can be found that the front car is too slow and wants to change lanes to the left to surpass the vehicle in front. After just changing lanes, it is found that it has reached the exit, then returned to the original lane, and then found the right time to transfer out. Facing the side-by-side exit, the Xiaomi YU7 smoothly drove into the correct position. Subsequent ramps are compact and have a large curvature. The Xiaomi YU7 began to gradually slow down before entering the curve, controlling the speed in the curve to about 30km/h to ensure the comfort of the car. Scenario 20: Autonomously entering the basement In the scenario of autonomously entering the basement, we have learned the route in advance. After setting up the navigation, the vehicle can seamlessly connect to the basement from the public road. Facing the sharp curve in the basement, Xiaomi YU7 will borrow the lane in advance. When the opposite car needs to be wrong, the vehicle will also react quickly, complete evasion, and then reverse to adjust the position. Eventually, you successfully park in a preset parking space, which is more efficient overall. Summarize the night peak after more than three hours of running, how will you evaluate the performance of Xiaomi YU7? At least my feeling is: bad reviews. The smart driving style of Xiaomi YU7 is a bit contradictory: smart, lively, but not calm enough and not focused enough. From the results, his performance was absolutely excellent. The takeover was basically harmless, and the scene pass rate was particularly high. In addition to turning around at three o'clock, basically all kinds of scenarios have been opened into the basement and out of the basement, past the toll station park, very smooth. However, on the other hand, YU7's intelligent driving also really inherited the track genes. Fast is really fast, and car sickness is also really easy to faint. Opening the method does not repair the margins, pursuing the limit, there are excessive small movements, the throttle brake is fierce, it seems impatient. In the future, it may be necessary to hone your personality more calmly and calmly in order to be more suitable for daily use. So Xiaomi YU7 smart driving like this, do you like it? Welcome to the comments section. This is the smart car studio that Zhou Zhi started broadcasting, we will see you next time!.