When people see movies about the future, they see robots that interact with their environment, learn quickly and adapt to changing conditions. Think of Chappie, Wall-E, I, Robot, Ex Machina.

We’ve all imagined a future where the environment we live in is interactive. You come home from work and don’t need to lock your car or your house. They lock and unlock themselves by recognizing you and your guests. You don’t need to order your own groceries because your fridge knows what you like and when it runs out. Your stove knows how to cook everything with the aid of a robot arm.

People work side-by-side with robots in factories, warehouses, mines, oil-fields and other unsafe environments, making work safer and increasing productivity. Fleets of self-driving cars enable people to socialize, work during their commute, and increase the safety and efficiency of our transport systems. Drones deliver packages to us increasing convenience, reducing emissions, and further freeing our roads and rails for transportation of people.

Much of this technology exists on its own – but none of it is in widespread use. These technologies span three similar areas – robotics, Internet of Things (IoT) and machine-to-machine communication (M2M) – all of which suffer from similar challenges to widespread adoption.

It is well-known that security and privacy are huge issues – and rightly so – nobody wants data about their travel, energy usage, food consumption or other identifying fingerprints exploited. Hacking autonomous vehicles has the potential for chaos, extreme loss of life and crippling the economies that would grow to depend on them.

In my opinion, however, this is not the biggest challenge. The biggest challenge is fragmentation and re-invention that goes on just bringing products to market in these areas. Whether it’s in robotics, home automation, connected devices, automated cars, forklifts and machinery – everybody starts from scratch.

The way we build connected things is all wrong. First we start with a computer. It can be something cheap and simple like an Arduino, Raspberry Pi, Beaglebone, or really high end – eight core Intel computer with dual video cards – and every specialized creation in between.

Here’s a closer look at four of the biggest challenges holding back robotics adoption.

Fragmented Platforms – One to Rule Them All?

The first issue is that none of these existing “platforms” were built to easily connect to a variety of parts that make interacting with environments easy. Some were built to make building hobby projects, like a connected plant, very easy to get going without a ton of expertise, while others were built to be fast video game computers, but also used for vision intensive robotics applications.

Some are low energy, but lack processing power. Some are powerful, but are heavy, hot, energy hogs that require large batteries.

Almost none are modular. None are standard. None are designed specifically for building devices that collect data from and interact with the environment.