Multi-sensor fusion is crucial for accurate and reliable autonomous driving systems. State-of-the-art approaches are based on point-level fusion: augmenting LiDAR point clouds with camera features. However, the camera-to-LiDAR projection discards the semantic density of camera features, hindering the effectiveness of such approaches, especially for semantic-oriented tasks such as 3D scene segmentation. In this paper, we propose BEVFusion, an efficient and general multi-task multi-sensor fusion framework. This approach unifies multimodal features in a shared bird's-eye view (BEV) representation space, effectively preserving both geometric and semantic information. To achieve this, we diagnose and eliminate key efficiency bottlenecks in view translation through optimized BEV pooling, reducing latency by over 40x. BEVFusion is fundamentally task-agnostic and seamlessly supports diverse 3D perception tasks with minimal architectural changes. It establishes a new state-of-the-art on the nuScenes benchmark, achieving 1.3% mAP and NDS for 3D object detection and 13.6% mIoU for BEV map segmentation, all while reducing computational cost by 1.9x.
A simple surround view system implementation, including the complete calibration, projection, stitching and real-time operation process. Detailed documentation can be found in the doc directory.
DCDCv9-3 is an electrically isolated DC/DC converter designed to convert 200V to 600V high voltage input to 24V low voltage output, with a continuous output power of up to 500W. Its credit card size (85.6×54mm) and 167g lightweight design are particularly suitable for powering automotive low voltage systems from high voltage batteries. Adopting resonant LLC soft switching topology, dynamic operating frequency (90-200kHz) ensures high efficiency and good EMC performance, and supports 12V to 48V output voltage customization.
The Speeduino project is a flexible, fully-featured engine management system (EMS aka ECU) based on the low-cost open-source Arduino platform. It provides the hardware, firmware, and software components that make up an engine management system, all under an open license. With over 1,000 installations, Speeduino has evolved into a product that meets the needs of the hobbyist and enthusiast community without the price tag of a traditional aftermarket ECU.
JetCar is a miniature self-driving car based on Jetson Nano. It can navigate on street maps and follow parking text and directional arrows.
CANable 2.0 is a small, low-cost, open-source USB to CAN adapter. CANable enumerates as a virtual serial port on your computer and acts as a serial line to CAN bus interface. Using the alternative candleLight firmware, CANable enumerates as a native CAN interface on Linux. CANable 2.0 supports standard CAN and CAN-FD.
The CANable adapter is compatible with ARM-based embedded platforms such as Raspberry Pi, Raspberry Pi Zero, ODROID, BeagleBone, etc., making it ideal for integration into OEM products.
WiCAN is a powerful CAN adapter based on ESP32-C3 that can be used for car hacking and general CAN bus development. It is available in two form factors: OBD-II and standard USB-CAN. The original firmware can interact directly with RealDash using Wi-Fi or BLE, which allows you to create custom dashboards with beautiful graphics. It is available for Android, iOS, and Windows 10. WiCAN connects to your existing Wi-Fi network and any device on that network, and it allows you to configure Wi-Fi and CAN settings through a built-in web interface. Both versions have a power saving mode that detects when the voltage drops below 13 V or other preset values. When this power saving mode is enabled, WiCAN is able to enter sleep mode, reducing the current consumption to less than 1 mA.
This is a fun electronic toy car that features power-ups! The author uses color sensors and magnetic switches to trigger fun tricks like acceleration, deceleration, and rotation. It draws a lot of inspiration from Mario Kart.
After a failed Kickstarter campaign, the author has now open-sourced the software, hardware, and mechanics.
A versatile LiDAR SLAM package. Supports various types of LiDAR sensors (mechanical, solid-state, etc.), 6-axis and 9-axis IMU, loose coupling and tight coupling, mapping and positioning, loop detection, etc.
Despite the growing demand for larger battery cells, battery prices remain quite high, constituting the most expensive component in an EV or PHEV, with a typical price tag of around $10,000 for a battery that supports a range of a few hundred kilometers. The high cost can be mitigated by using lower-cost/refurbished battery cells, but such cells will also have greater capacity mismatches, which will reduce the available runtime or driving distance on a single charge. Even higher-cost, higher-quality battery cells will age and mismatch after repeated use. There are two ways to increase the capacity of a battery pack with mismatched cells: one is to use larger batteries from the beginning, which is not cost-effective; the other is to use active balancing, a new technology that can restore battery capacity in the battery pack and quickly increase power. Full series battery cells need balancing When each battery cell in the battery pack has the same state of charge (SoC), we say that the battery pack is balanced. SoC refers to the current remaining capacity of an individual battery relative to its maximum capacity as the battery is charged and discharged. For example, a 10Ah battery will automatically equalize the state of charge between parallel-connected battery cells over time as long as there is a conductive path between the battery cell terminals. It can also be argued that the state of charge of series-connected cells will vary over time for a variety of reasons. Temperature gradients across the pack, impedance, self-discharge rates, or differences in load between individual cells can cause gradual changes in SoC. While pack charge and discharge currents help to minimize these cell-to-cell differences, the cumulative mismatch will only increase unless the cells are periodically balanced. Compensating for gradual changes in cell SoC is the most fundamental reason to balance series-connected cells. Typically, passive or dissipative balancing schemes are sufficient to rebalance the SoC of cells with similar capacities in the pack.
The reference design is a BLDC motor controller designed to be powered by a single 12V (nominal voltage) supply with a wide voltage range found in typical automotive applications. The board is designed to drive motors in the 60W range, which requires a current of 5 amps. The size and layout of the board facilitates evaluation of the drive electronics and firmware, with easy access to key signals on various test points. A wide variety of motors can be connected by using a 3-contact connector or soldering the motor phase wires to the plated through holes in the board. The 12VDC supply is fused to prevent damage to the board or bench power supply in the event of a motor failure during testing. Commands and the status of the motor can be transmitted through a standard JTAG connector or through PWM input and output signals. The user can also reprogram the microcontroller through the JTAG connector, allowing customization for various applications. This design forms the solution by incorporating the DRV8301-HC-C2-KIT board.
With features like adaptive cruise control, driver monitoring, automatic lane centering, etc., it is available for Toyota, Hyundai, Honda and many other brands, about 275+ models. openpilot complies with ISO26262 guidelines.
Typical Application of NCV890203 2.0 A, 2 MHz Automotive Step-Down Switching Regulator with RSTB. The NCV890203 is a fixed frequency, monolithic, step-down switching regulator designed for automotive battery-connected applications that must operate from an input supply up to 36 V.
EEWorld
subscription
account
EEWorld
service
account
Automotive
development
community
Robot
development
community
About Us Customer Service Contact Information Datasheet Sitemap LatestNews
Room 1530, 15th Floor, Building B,
No.18 Zhongguancun Street,
Haidian District,
Beijing, Postal Code: 100190
China
Telephone: 008610 8235 0740
京公网安备 11010802033920号
