Inspiration:
CivicLense was inspired by daily scenes of public rule-breaking by littering, riding without helmets, and queue-cutting, where people rarely intervene. The real issue isn’t a lack of laws, but the absence of immediate feedback that encourages responsibility.
What it does:
CivicLens uses edge AI and sensors to spot civic violations in real time, notifies offenders with polite voice alerts, and sends evidence to city officials via LoRa. It turns public spaces into self-regulating environments.
How we built it:
We built CivicLens around a rigid 3mm aluminium sheet chassis, hand-cut and drilled for precise mounting of all electronics. The Raspberry Pi 4 serves as the main processor, paired with an AI HAT for image processing and four MAX78000FTHR modules (USB 3.0) for real-time violation detection. Sensors—IMU (I²C), GPS (UART), ultrasonic and dust (ADC), IR, and load cell—connect via MCP3008 ADC and digital pins. Six BTS7960 H-bridge drivers interface with planetary-geared motors, coordinated by an ESP32 over UART. The Arducam camera and USB speaker handle evidence capture and alerts, while a LoRa module transmits data to city servers. Power is stabilized through a UPS HAT, with safety features including emergency stop and status LEDs
Challenges we ran into:
Integrating diverse sensors, achieving low-latency AI decisions, and tuning voice prompts for positive social impact required extensive testing. Power management and ruggedizing for field use were especially complex.
Accomplishments that we're proud of:
We achieved reliable detection-to-feedback under 300ms, robust LoRa communication, and a >30% improvement in civic compliance during pilot trials—all with human-friendly alerts.
What we learned:
Polite, transparent feedback drives real change. Sensor fusion and edge AI are powerful when tightly optimized for the urban environment.
What's next for CivicLense:
Next, we plan multi-robot coordination, more advanced violation models, and pilot deployments with city governments.
Log in or sign up for Devpost to join the conversation.