Smart Sorting Systems – 5 concepts (IEEE IES Day Jordan 2026)

Assumption: small 2.5–4 cm objects (balls, cubes, caps) sorted mainly by colour, optionally shape/height/metal. The competition rules were not provided; adapt dimensions and classes once they are known.

New: interactive working simulations. All five concepts are now fully working simulations - open any from its section below.

What makes a system stand out

• Complete pipeline that visibly works end to end: detect → classify → decide → move.
• Reliability beats complexity: judges reward a system that sorts 20/20 objects.
• One clear original idea, explainable in 10 seconds.
• Visual impact: transparent parts, moving mechanisms, a live camera view on screen.
• Buildable by students in weeks with acrylic, 2020 profile, plywood and 3D prints.

Scorecard (1–10, my estimate)

Recommendation: Revolver Vault (best reliability + originality) as the main entry; TiltSort Arena if time or budget is tight.

01 · Orbit Picker

Radial 3-DOF robotic arm picks each object from a pad and places it in a ring of bins.

▶ Open interactive simulation

How it works

IR break-beam detects the object on the pad → LED ring + camera trigger → Raspberry Pi (OpenCV HSV + contour shape) classifies → class→bin table + inverse kinematics on Arduino Mega → servos grip, swing, release, home.

Sensing & classification

Camera, IR break-beam, LED ring; OpenCV colour/shape

Sorting mechanism

Base yaw servo + shoulder/elbow servos + gripper servo

Main hardware

Pi 4, Arduino Mega, PCA9685, 3×MG996R + 1 micro servo, 5 V 10 A PSU, 2020 profile mast, plywood base, 4 bins

Key innovation

Any visual rule can be coded with no mechanical change; most “industrial” demo.

Build complexity

Medium-high: arm calibration and gripper tuning take the most time.

02 · Revolver Vault

Object drops into a lit inspection chamber; a stepper-indexed rotating chute delivers it to one of 5 cups.

▶ Open interactive simulation

How it works

IR gate under hopper → object lands on trapdoor → camera + RGB sensor vote (disagree = reject) → stepper rotates chute to the cup angle → servo opens trapdoor → gravity slide.

Sensing & classification

Camera + TCS34725-type RGB sensor, IR gate

Sorting mechanism

NEMA 17 + GT2 belt rotating ring; trapdoor servo

Main hardware

Uno, Pi 4, stepper driver, NEMA 17, micro servo, acrylic chamber, 5 cups, aluminium posts

Key innovation

Bins scale by adding cups; dual-sensor agreement makes it very reliable.

Build complexity

Medium: only one real mechanism (rotary ring), little to jam.

03 · TiltSort Arena

Two servos tilt a clear acrylic plate; the object rolls off the correct edge into a bin.

▶ Open interactive simulation

How it works

IR gate sees the drop → overhead camera tracks colour and live (x,y) → class→edge → PD loop sets tilt angles → plate tilts, object slides into bin, plate levels.

Sensing & classification

Overhead camera, LED ring, IR gate

Sorting mechanism

2 servos + push-rods on a ball-joint pivot

Main hardware

Uno, Pi 4, PCA9685, 2 servos, 28 cm acrylic plate, ball joint, 4 bins, L-gantry

Key innovation

The platform is the actuator: no belt, arm or gripper. Cheapest, most dramatic demo.

Build complexity

Low-medium: needs low-friction plate and servo tuning; keep objects round or smooth.

04 · TriSense Line

Conveyor with vision, height and metal sensing; servo paddles push objects into bins.

▶ Open interactive simulation

How it works

IR beam at entry → tunnel scan (camera colour/shape + ToF height + inductive metal) fused into one class → belt-speed timing picks the pusher → paddle pushes; unknown rides to reject bin.

Sensing & classification

Camera, VL53L0X ToF, inductive sensor, IR beam

Sorting mechanism

DC gear motor belt + 3 servo pushers

Main hardware

Mega, Pi 4, L298N, gear motor, 3 servos, belt + rollers, 2020 profile, 4 bins

Key innovation

Multi-attribute sorting (colour + size + material) on the most practical architecture.

Build complexity

Medium: belt tracking and pusher timing are the fiddly parts.

05 · Cascade Tower

Archimedes-screw lift raises objects to a scan chamber; a servo flap tree sorts them by gravity into 4 bins.

▶ Open interactive simulation

How it works

IR gate starts the lift motor → screw carries object up → chamber gate holds it while camera + RGB classify → 2-bit code sets flaps A and B1/B2 → gate opens, object falls into its bin.

Sensing & classification

Camera, RGB sensor, IR gate

Sorting mechanism

DC gear motor auger; gate servo; 3 flap servos

Main hardware

Mega, Pi 4, L298N, PCA9685, gear motor, 4 servos, acrylic channel, 3D-printed auger

Key innovation

Zero-contact sorting through a transparent Plinko-style tree; very photogenic.

Build complexity

High: auger fit/jamming and tall frame are the risks.

06 · Pustache Sorter (upgraded Tinkercad design)

Your Tinkercad tower made competition-ready: feed tube + two-pin escapement (one object at a time), lit inspection chamber with IR beam and camera, HSV vote with a reject cup, servo swing chute over four cups.

▶ Open interactive simulation

⬡ Open 3D model (explode, parts list, STL download)

How it works

Objects queue in the tube → escapement releases exactly one → IR beam confirms entry → LED + camera capture 5 frames → majority vote and confidence check → unsure = one re-scan, then REJECT → Arduino aims the chute servo, opens gate G1 → object slides into the cup → IR drop-confirm.

Sensing & classification

Pi camera + OpenCV HSV vote, IR break-beam, IR drop-confirm per cup

Sorting mechanism

One servo swing chute over 4 cups on an arc; escapement servo; gate servo

Main hardware

Pi 4, Arduino Uno, 3 servos, LED strip, 5 IR pairs, camera, existing tube/shelves/cups

Key innovation

Reject path + teach mode for venue lighting + closed-loop drop confirmation.

Build complexity

Low-medium: three servos and simple printed parts; reuses your existing frame.