Self-Driving Cars with Fexingo: Autonomous Vehicles, Lidar, and Mobility Tech

In episode 54 of Self-Driving Cars with Fexingo, Lucas and Luna dig into why autonomous vehicles struggle with unpaved roads. Using recent data from Rivian and Lucid, they explore how gravel, dirt, and loose surfaces confuse sensor suites and control algorithms. The conversation covers lidar limitations, traction modeling failures, and why most AV testing ignores rural routes. Specific examples in...

In this episode of Self-Driving Cars with Fexingo, Lucas and Luna dive into one of the most stubborn blind spots for autonomous vehicles: tunnels. From GPS dropouts and lidar bounce to the deadly 2023 Tesla crash in the Bay Area, they explain why even the most advanced robotaxis struggle underground. They discuss sensor fusion challenges, how automakers like Tesla and Waymo are tackling the proble...

Self-driving cars have been tested for years, yet construction zones remain a major stumbling block. In this episode, Lucas and Luna explore why temporary lane shifts, ambiguous signage, and human flaggers confuse even the most advanced sensor stacks. They discuss how Tesla's vision-only approach struggles where LiDAR excels, and why mapping companies like Waymo and Mobileye are investing in real-...

Episode 51 of Self-Driving Cars with Fexingo tackles a bizarre but critical edge case: funeral processions. Lucas and Luna break down why autonomous vehicles struggle with the convoy behavior, right-of-way violations, and social norms of funeral processions. They reference recent market moves in Rivian and Lucid, and explain how a single procession in Ohio last year exposed fundamental gaps in AV...

Episode 50 of Self-Driving Cars with Fexingo explores a persistent blind spot in autonomous vehicle perception: animals. From deer on rural roads to dogs darting into traffic, AVs struggle to detect, classify, and predict the erratic movement of animals. Lucas and Luna discuss the technical challenge—why lidar and cameras miss wildlife, how machine learning models fail on rare animal shapes and be...

In Episode 49 of Self-Driving Cars with Fexingo, Lucas and Luna tackle a persistent blind spot for autonomous vehicles: pedestrians who cross outside crosswalks. With jaywalking laws being decriminalized in cities like New York and Denver, self-driving cars face a new challenge—predicting erratic human behavior where no path is defined. The hosts dig into how Waymo, Cruise, and others train for th...

In this episode of Self-Driving Cars with Fexingo, Lucas and Luna dive into why roundabouts remain a persistent challenge for autonomous vehicles. They discuss the technical hurdles of predicting human driver intent at yield signs, the complex geometry of multi-lane circles, and how even Waymo's fleet in San Francisco struggles with these intersections. With roundabouts growing in popularity acros...

In this episode, Lucas and Luna explore why self-driving cars still struggle to safely navigate around construction workers. Using data from a 2025 study by the University of Michigan, they discuss how autonomous vehicles fail to predict the erratic movements of workers on foot near heavy machinery. They also touch on recent stock performance of companies like Tesla and GM, and how Mercedes-Benz i...

Episode 46 of Self-Driving Cars with Fexingo examines why potholes remain a stubborn challenge for autonomous vehicle perception systems. Lucas and Luna break down how potholes break lidar and camera fusion, why they confuse neural nets trained on smooth roads, and how companies like Waymo and Tesla approach the problem differently. They also discuss real-world failures, the role of road maintenan...

In this episode of Self-Driving Cars with Fexingo, Lucas and Luna dig into one of the most stubborn challenges for autonomous vehicles: snow. With winter conditions covering roads in snow, ice, and slush, lidar and camera systems struggle to detect lane markings, curbs, and other vehicles. The hosts discuss how companies like Waymo and Tesla approach the problem, why simulation alone isn't enough,...

Lucas and Luna dive into a little-discussed failure mode of autonomous vehicles: nighttime operation. Despite lidar and radar, self-driving systems struggle with low-light conditions, glare from oncoming headlights, and unexpected darkness. The episode focuses on a 2025 study from the University of Michigan showing that AVs have a 5.1 times higher disengagement rate at night compared to daylight....

Lucas and Luna examine why heavy rain, fog, and snow remain a major unsolved challenge for autonomous vehicles. They discuss how lidar and camera performance degrades in precipitation, why current sensor fusion struggles with wet-road reflections and reduced visibility, and what companies like Waymo and Tesla are doing to mitigate weather-related failures. Luna brings data from a University of Mic...

Lucas and Luna dig into one of the most stubborn blind spots in autonomous driving: unpaved roads. Most self-driving systems are trained exclusively on high-definition maps of paved streets. When a vehicle hits gravel, packed dirt, or mud, its sensors lose reference points, lane markings vanish, and the AI struggles to interpret traction. Lucas cites a recent study from the University of Michigan...

Episode 41 of Self-Driving Cars with Fexingo digs into a persistent blind spot: highway merging. Lucas and Luna break down why autonomous vehicles freeze or panic at on-ramps, using real-world data from California DMV disengagement reports. They explore the geometry problem—merging requires predicting human drivers' intentions at high speed—and how companies like Waymo and Tesla approach it differ...

Episode 40 of Self-Driving Cars with Fexingo digs into a stubborn edge case: unpaved roads. Lucas and Luna explore why lidar and cameras fail on gravel, dust, and loose surfaces, and how companies like Waymo and Tesla are tackling the problem. They reference Rivian's recent 2.6% drop amid off-road autonomy challenges and GM's 2.7% gain as Cruise revisits rural testing. The conversation covers sens...

Episode 39 of Self-Driving Cars with Fexingo tackles a blind spot in autonomous vehicle perception: motorcycles. Lucas and Luna break down why lidar and camera systems struggle to track two-wheeled vehicles, citing a recent IIHS study showing that AVs detect motorcycles 30% less reliably than cars. They explore the physics of narrow profiles, the lack of training data, and what Waymo and Tesla are...

Episode 38 of Self-Driving Cars with Fexingo tackles one of the most dangerous edge cases for autonomous vehicles: emergency vehicles. Lucas and Luna dig into new data from California's autonomous vehicle disengagement reports, which show that emergency vehicle encounters remain a top reason for human takeovers. They discuss why sirens and flashing lights confuse AI perception systems, how compani...

Episode 37 of Self-Driving Cars with Fexingo digs into a stubborn blind spot for autonomous vehicles: school zones. Lucas and Luna explore why robotaxis still struggle with the chaotic mix of crossing guards, flashing signs, and unpredictable kids. They cite recent data showing a 23.1% drop in Lucid stock and a 10.4% slide in Ford, linking investor skepticism to the industry's failure to solve edg...

Episode 36 of Self-Driving Cars with Fexingo tackles the surprisingly stubborn problem of autonomous parking. While robotaxis handle complex highway merges and unprotected left turns, parallel parking and crowded garages still trip them up. Lucas and Luna break down why sensor blind spots, dynamic environments, and edge cases like valet lots remain unsolved. They reference recent stock drops in Te...

In this episode, Lucas and Luna dive into a surprisingly tricky corner of autonomous driving: traffic lights. While most people assume traffic signals are simple for self-driving cars, the reality is far messier. Lucas explains how different municipalities use wildly different hardware and timing patterns — from old-school incandescent bulbs to LED arrays with strobe-like refresh rates. He breaks...

Lucas and Luna dive into one of the most deceptively hard problems in self-driving: the four-way stop sign. Despite billions in R&D and millions of miles driven, autonomous vehicles still struggle with the subtle social negotiation of who goes next. Lucas breaks down why this isn't just a software bug — it's a fundamental challenge of predicting human intention. He cites new data from Waymo's Phoe...

Lucas and Luna explore a persistent blind spot in self-driving technology: the inability to reliably detect and respond to emergency vehicles like ambulances and fire trucks. They examine real-world incidents, including an April 2026 crash involving a Waymo vehicle in Phoenix, and discuss why siren recognition and light detection remain harder than lidar mapping. The episode touches on training da...

Self-driving cars can handle highways and city streets, but roundabouts remain a persistent challenge. In this episode, Lucas and Luna examine why circular intersections break autonomous vehicle logic — from unpredictable human drivers to occlusion problems and map limitations. They discuss how companies like Waymo and Cruise are testing new approaches, including dedicated roundabout training data...

Unprotected left turns remain one of the hardest unsolved problems for self-driving cars. In this episode, Lucas and Luna unpack why a single human maneuver that takes two seconds is still stumping autonomous systems with billions of miles of training data. They walk through the sensor fusion challenge: how lidar, radar, and cameras disagree on the speed of an oncoming car at a 30-degree angle. Th...

Uber announced plans to put 500 data-collection vehicles on the road this year. Lucas and Luna dig into why this fleet exists, what it tells us about the state of autonomous vehicle training data, and how Uber is playing catch-up with Waymo and Tesla. They discuss the shift from simulation-only training to real-world data loops, the economics of running a dedicated mapping fleet, and what this mea...